Friedel Drees Motives for and Consequences of Minority Equity Purchases
GABLER RESEARCH EBS Forschung Schriftenreihe ...
10 downloads
425 Views
628KB Size
Report
This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!
Report copyright / DMCA form
Friedel Drees Motives for and Consequences of Minority Equity Purchases
GABLER RESEARCH EBS Forschung Schriftenreihe der European Business School (EBS) International University · Schloss Reichartshausen Herausgegeben von Univ.-Prof. Ansgar Richter, PhD
Band 76
Die European Business School (EBS) – gegründet im Jahr 1971 – ist Deutschlands älteste private Wissenschaftliche Hochschule für Betriebswirtschaftslehre im Universitätsrang. Dieser Vorreiterrolle fühlen sich ihre Professoren und Doktoranden in Forschung und Lehre verpflichtet. Mit der Schriftenreihe präsentiert die European Business School (EBS) ausgewählte Ergebnisse ihrer betriebs- und volkswirtschaftlichen Forschung.
Friedel Drees
Motives for and Consequences of Minority Equity Purchases With a foreword by Prof. Dr. Dirk Schiereck
RESEARCH
Bibliographic information published by the Deutsche Nationalbibliothek The Deutsche Nationalbibliothek lists this publication in the Deutsche Nationalbibliografie; detailed bibliographic data are available in the Internet at http://dnb.d-nb.de.
Dissertation European Business School, International University Schloss Reichartshausen, Oestrich-Winkel, 2009 D 1540
1st Edition 2010 All rights reserved © Gabler Verlag | Springer Fachmedien Wiesbaden GmbH 2010 Editorial Office: Ute Wrasmann | Britta Göhrisch-Radmacher Gabler Verlag is a brand of Springer Fachmedien. Springer Fachmedien is part of Springer Science+Business Media. www.gabler.de No part of this publication may be reproduced, stored in a retrieval system or transmitted, in any form or by any means, electronic, mechanical, photocopying, recording, or otherwise, without the prior written permission of the copyright holder. Registered and/or industrial names, trade names, trade descriptions etc. cited in this publication are part of the law for trade-mark protection and may not be used free in any form or by any means even if this is not specifically marked. Umschlaggestaltung: KünkelLopka Medienentwicklung, Heidelberg Printed on acid-free paper Printed in Germany ISBN 978-3-8349-2164-2
V
Foreword Share price reactions to the announcement of (partial) corporate acquisitions have extensively been analysed in capital market research, in particular in the context of event studies. Minority equity purchases have also been examined before, but never as comprehensively or with a special focus on the underlying explanatory variables as in this doctoral thesis. Moreover, previous empirical research has largely focused on the U.S. capital market, which shows significant differences to most European countries with regard to the regulatory framework of minority shareholder protection. Consequently, the transatlantic evidence can only be transferred to the institutional and regulatory framework in Europe to a very limited extent. There are only very few extensive insights for the German capital market and comparative studies in a European context. It is all the more welcome that this doctoral thesis has not only focused on the U.S., but also on German and European markets. It thus allows differentiated insights into highly distinct systems of corporate governance. This doctoral thesis addresses an important research gap and stands out due to its special attention to detail and accuracy. Its primary objective is the empirical explanation of wealth effects in corporations with new minority equity holders. The objective insights derived in this thesis will serve as a well-grounded basis for the decision making of institutional investors, corporate practitioners, and M&A professionals. In this doctoral thesis, Mr. Drees has optimally achieved all the set objectives – his work contains a great number of highly interesting results and is written in a way which the reader will find very enjoyable. I wish his thesis the wide circulation it deserves.
Professor Dr. Dirk Schiereck
VII
Acknowledgements My deepest gratitude goes to my advisor Dirk Schiereck for his mentoring and support during the last 2 years. His detailed and constructive feedback helped me to shape my research ideas and to keep me on track. For his support and encouragement – academically and personally – I would like to express my sincere gratitude. It was a pleasure being grated the opportunity to work with you! I also thank Marcel Tyrell for accepting the role as a co-advisor. He provided invaluable ideas and feedback, especially with regard to the modelling and economic interpretation of the concept of “corporate relatedness” which I apply in the last chapter of my dissertation. Without naming each individual, I am grateful to the various other supporters of my thesis. I especially appreciated the feedback and comments from my fellow PhD students, from my former colleagues and friends at Bain & Company, and from numerous anonymous reviewers at journals. Last but not least, I would like to thank my beloved parents, Franz and Beatrix, who always provided me with their unequivocal support and gifted me with the traits and education that enabled me to accomplish this academic endeavour.
Friedel Drees
IX
Table of Contents List of Tables & Figures ............................................................................................... XI List of Abbreviations ................................................................................................. XIII List of Symbols ............................................................................................................. XV 1
Introduction .............................................................................................................. 1
2
New Outside Blockholders, Performance, and Governance in Germany ........... 9
2.1 Introduction ............................................................................................................ 9 2.2 Literature Review ................................................................................................. 11 2.2.1 General Considerations with Regard to the Potential Wealth Effects of Block Purchases ............................................................................ 11 2.2.2 Empirical Evidence on The Effects of Block Purchases in Germany ........... 13 2.3 Data and Methodology ......................................................................................... 14 2.3.1 Sample Selection ........................................................................................... 14 2.3.2 Block Characteristics ..................................................................................... 17 2.3.3 Target Characteristics .................................................................................... 20 2.3.4 Method of Measuring Abnormal Returns ...................................................... 22 2.4 Results .................................................................................................................. 22 2.4.1 Announcement Effects................................................................................... 22 2.4.2 Determinants of Abnormal Returns ............................................................... 25 2.5 Conclusion and Outlook ....................................................................................... 28 3
Equity Ownership, Agency Problems and Shareholder Wealth: Understanding the Unique Role of Corporate Block Owners ............................ 31
3.1 Introduction .......................................................................................................... 31 3.2 Literature Review ................................................................................................. 34 3.2.1 Theoretical Benefits of Corporate Block Equity Ownership ......................... 34 3.2.2 Empirical Evidence on the Wealth Effects of Corporate Minority Block Purchases ............................................................................................. 36 3.3 Data and Methodology ......................................................................................... 38 3.3.1 Corporate Block Purchases ............................................................................ 38 3.3.2 Factors that Explain the Value Creation from Corporate Block Equity Purchases ............................................................................................ 42 3.3.2.1 Industry Characteristics ..................................................................... 44
X
3.3.2.2 Product Market Relationship Related Characteristics....................... 46 3.3.2.3 Transaction Characteristics ............................................................... 47 3.3.2.4
Target Financial Characteristics ........................................................ 49
3.3.3 Method of Measuring Abnormal Returns ...................................................... 52 3.4 Results .................................................................................................................. 53 3.4.1 Univariate Analysis ....................................................................................... 53 3.4.2 Cross-Sectional Regression Analysis ............................................................ 59 3.5 Conclusions .......................................................................................................... 65 4
Minority Equity Ownership and Value Creation: The Role of Corporate Relatedness ...................................................................... 67
4.1 Introduction .......................................................................................................... 67 4.2 Hypotheses and Related Literature ...................................................................... 70 4.3 Data and Methodology ......................................................................................... 72 4.3.1 Corporate Block Purchases ............................................................................ 72 4.3.2 Measuring Corporate Relatedness ................................................................. 76 4.3.2.1 Benchmark Input-Output Accounts .................................................. 76 4.3.2.2 Method of Measuring Vertical Relatedness ...................................... 77 4.3.2.3 Method of Measuring Complementarity ........................................... 80 4.3.3 Method of Measuring Abnormal Returns ...................................................... 82 4.4 Results .................................................................................................................. 82 4.4.1 Univariate Analysis ....................................................................................... 82 4.4.2 Cross-Sectional Regression Analysis ............................................................ 86 4.4.3 The Effect of Secondary Segments................................................................ 91 4.5 Conclusion ............................................................................................................ 96 5
Concluding Remarks .............................................................................................. 99
References .................................................................................................................... 103
XI
List of Tables & Figures Table 2–1: Sample Selection .......................................................................................... 16 Table 2–2: Distribution of Transactions by Year and Block Type ................................. 16 Table 2–3: Sample Description of Block Characteristics ............................................... 19 Table 2–4: Sample Description of Target Characteristics .............................................. 21 Table 2–5: Excess Returns by Acquiring Block Type.................................................... 23 Table 2–6: Determinants of Cumulative Average Abnormal Target Returns ................ 26 Table 3–1: Sample Selection .......................................................................................... 40 Table 3–2: Distribution of Block Purchases by Year ..................................................... 41 Table 3–3: Sample Transactions by Home Country of Target and Acquirer ................. 41 Table 3–4: Distribution of Acquiring and Target Firms by Industry and Year .............. 42 Table 3–5: Variable Description ..................................................................................... 43 Table 3–6: Descriptive Statistics of Continuous and Binary Variables ......................... 51 Table 3–7: Excess Returns around Block Equity Purchases by Corporations ............... 55 Table 3–8: Determinants of Cumulative Average Abnormal Target Returns ................ 60 Table 3–9: Determinants of Cumulative Average Abnormal Acquirer Returns ............ 64 Table 4–1: Sample Selection .......................................................................................... 75 Table 4–2: Distribution of Block Purchases by Year ..................................................... 75 Table 4–3: Distribution of Block Purchases by Industry................................................ 76 Table 4–4: Calculation of Vertical Relatedness and Complementarity Coefficients ..... 79 Table 4–5: Sample Description ...................................................................................... 81 Table 4–6: Excess Returns Around Block Equity Purchases by Corporations .............. 83 Table 4–7: Mean Difference Test of Excess Returns ..................................................... 85 Table 4–8: Variable and Source Description .................................................................. 8 Table 4–9: Determinants of Cumulative Average Abnormal Combined Entity Returns based on Primary Industry Classification ............................ 8 Table 4–10: Descriptive Statistics of Control Sample .................................................... 92 Table 4–11: Excess Returns around Block Purchases for Control Sample .................... 93 Table 4–12: Mean Difference Test of Excess Returns for Control Sample ................... 94
XII
Table 4–13: Determinants of Cumulative Average Abnormal Combined Entity Returns for Control Sample ............................................................. 95 Figure 2–1: Cumulative Average Abnormal Target Returns ......................................... 24 Figure 3–1: Cumulative Average Abnormal Returns ..................................................... 56
XIII
List of Abbreviations AktG
Companies Act
AR
Abnormal return
BaFin
German Federal Financial Supervisory Authority
CAAR
Cumulative average abnormal return
CDAX
Composite DAX (German equity index)
ESA
European System of Accounts
M&A
Mergers and acquisitions
IO table
Input-output table
R&D
Research and development
WpHG
Securities Trading Act
SIC
Standard Industrial Classification
XV
List of Symbols aij
Dollar value of industry i’s output required to produce industry j’s total output
bik
Percentage of industry i’s output supplied to intermediate industry k
e1,e2
First day, last day in event window
CARi ,e1 e2
Cumulated abnormal return of stock i in event window [e1, e2]
Qi
Total output of industry i
RM,t
Return on market index on day t (event period)
RM
Average return on market index in estimation period
sˆi
Estimated standard deviation of abnormal returns for stock
SRi,t
Standardized abnormal return on stock i at day t
t0
Event date
R
2
Coefficient of determination
T
Number of days in estimation period
Ts
Number of days in event window
vij
Dollar value of industry i’s output necessary to produce one dollar’s worth of industry j’s output
vki
Percentage of intermediate industry k’s output supplied to industry i
1
1
Introduction
The notion of diffuse stock ownership and its inherent problems is well entrenched in financial research. It has already been addressed by early economists such as Smith (1776) and Berle and Means (1932).The latter state: “[…] the dissolution of the old atom of ownership into its parts, control and beneficial ownership […] destroys the very foundation on which the economic order of the past centuries has rested.” In their work on the agency problems that ultimately stem from this dissolution, Jensen and Meckling (1976) develop a more complete view on the problem of diffuse stock ownership. They integrate elements from the theory of agency, the theory of property rights, and the theory of finance to develop a new theory of the ownership structure of the firm, which focuses on the conflict between diffuse shareholders and professional managers: “Since the relationship between the stockholders and managers of a corporation fit the definition of a pure agency relationship, it should be no surprise to discover that the issues associated with the “separation of ownership and control” in the modern diffuse ownership corporation are intimately associated with general problems of agency. We show […] that an explanation of why and how the agency costs generated by the corporate form are born leads to a theory of the ownership (or capital) structure of the firm.” Since the seminal work of Shleifer and Vishny (1986), it has been well entrenched in financial research that large outside blockholders theoretically can play an important role in mitigating these agency problems by behaving as monitors. This notion gave rise to a strand of empirical research that employed the agency perspective to examine in more detail the specific role of large blockholders. For example, Demsetz and Lehn (1985) analyze the question of the types of public corporations that are likely to have high managerial ownership. Holderness and Sheehan (1988) examine whether important corporate decisions are influenced by the presence of a large-percentage shareholder. Morck,
2
Shleifer and Vishny (1988) address the question whether different levels of managerial ownership influence firm value. Despite the broad body of literature that addresses the role of large blockholders, there are still numerous research gaps with regard to the specific motives, characteristics and effects of different types of large blockholders. Holderness (2003) in his literature review on large blockholders and corporate control states: “Perhaps above all, the academic literature highlights the richness of blockholders. An outsider blockholder, for instance, has a different set of incentives than does a CEO blockholder. Blockholders have the incentive to improve management, but they also have the incentive to consume corporate resources. Blockholders that are corporations present a set of issues not found with those who are individuals. Because of this richness, the literature on blockholders and corporate control will continue to grow, and with it our understanding of the modern public corporation will deepen.” Accordingly, Cronqvist and Fahlenbrach (2007) argue that the incorporation of “blockholder heterogeneity” is highly relevant for the economic analysis of large shareholders. However, they criticize that “blockholder heterogeneity” is not incorporated in existing empirical frameworks and show that large blockholders have different investment and governance preferences and conducts. These considerations with regard to the economic relevance of blockholder identity and characteristics are the starting point for my doctoral thesis which consists of three papers. By analyzing the effects of corporate block purchases on firm value in different contexts, the papers of this thesis contribute to the understanding of (minority) blockholders. In the following, I briefly discuss the key research questions that are analyzed in this thesis.
First research question: Although there is some empirical support for the notion that large blockholders can help to limit agency problems through the involvement in monitoring or control activities
3
(e.g. Agrawal and Mandelker (1990), Hartzell and Starks (2003)), results are ambiguous (Holderness (2003)). Most of the existing studies are static in nature, examining the impact of already existing large blockholders on firm performance or monitoring related variables such as manageement turnover. In contrast, the body of empirical research on the dynamic effects of changes in ownership structure is considerably smaller. This strand of literature examines acquisitions of 5% or more, that is, the emergence of new blocks (e.g. Choi (1991), Barclay and Holderness (1991), Bethel, Liebeskind and Opler (1998), Allen and Phillips (2000), Park, Selvili and Song (2008)). So far, block purchases and the determinants of the concurrent wealth effects appear relatively well explored. However, previous research is predominantly focused on the U.S. In contrast, evidence on the role and effect of large blockholders in Europe is surprisingly scarce, despite significant differences in the system of corporate control and governance. Hence, we want to fill this gap by empirically analyzing the effect of block purchases in Germany. The German system of corporate control and governance has several unique characteristics that motivate us to specifically focus on block purchases in Germany. First, Germany in particular is characterized by a significantly higher ownership concentration than the U.S. (La Porta, Lopez-de-Silanes and Shleifer (1999), Faccio and Lang (2002)). Becht and Boehmer (2003) document that over 80% of officially exchange listed German corporations have a minority blockholder controlling over 25% of the votes and that 65% are majority controlled. In contrast to the U.S., a considerable power can be associated with blocking minority stakes (larger than 25%), even in cases where a controlling shareholder or a coalition of controlling shareholders exists (Jenkinson and Ljungquist (2001)). Second, large stakes have a significantly higher relevance to the market for corporate control than in the U.S. Franks and Mayer (2001) even regard them as “[…] a substitute for the traditional market of corporate control”, as they can lead to both friendly and hostile control changes (Jenkinson and Ljungqvist (2001)). Given the above considerations, we analyze a new set of data with 85 block purchases with German targets between 1997 and 2006. We can document a significant
4
wealth creation to target companies, regardless the economic intention of the acquirer. Most importantly, this finding is consistent with the U.S. evidence, despite the structural differences discussed above. Focusing on cross-sectional block and target characteristics such as acquirer intention and ownership structure, we can confirm a significant positive relationship for specific types of acquirers. Our results also suggest that existing ownership structure has at most a marginal impact on the market reaction. Taking into account the potential impact of the acquirer’s identity on the capital markets’ reaction, we want to further explore the causes and consequences of block purchases for a specific type of acquirer: non-financial corporations.
Second research question: Non-financial corporations are frequent acquirers and owners of large equity blocks in other non-financial corporations in both Europe (Dherment-Ferere, Köke and Renneboog (2004), Kirchmaier and Grand (2005)) and the U.S. (Cronqvist and Fahlenbrach (2007)). In contrast to other, well-studied types of blockholders such as banks or other institutional investors, corporations have several unique characteristics that motivate a more differentiated analysis. For example, business agreements, alliances or joint ventures may be formed between target firms and corporate owners. Despite these unique characteristics, there is still a substantial research gap with regard to the motives and effects of corporate block ownership. Accordingly, Allen and Phillips (2000) state: “[…] despite the fact that corporate block ownership is potentially quiet different from block ownership by institutional owners, we know very little about equity holdings by non-financial corporations.” Although the causes and effects of corporate block ownership have recently been addressed more thoroughly (Allen and Phillips (2000), Fee, Hadlock and Thomas (2006)), there is still no consistent and complete empirical insight into the specific sources of value creation in corporate block purchases. To address this research gap, we analyze a sample of 113 European minority equity purchases by non-financial corporations during the 1993 to 2006 period. We examine the
5
cumulative abnormal returns to target firms, acquiring firms and the value-weighted combined returns around the announcement of the corporate block purchase. We then examine whether cross-sectional differences in excess returns can be explained based on product market related characteristics, industry characteristics, transaction characteristics and target financial characteristics. We document a significantly positive stock price reaction for the target firms at the announcement of the block purchase, indicating a significant wealth creation for target shareholders. Moreover, we find that the combined value-weighted returns are positive and significant, which suggests that the formation of a corporate minority equity stake adds economic value overall. We also can exclude systematic wealth transfers from the purchasing firms to the target firms and vice versa. Considering the factors that relate to the specific hypotheses regarding the potential benefits of corporate block purchases such as the alignment of incentives or the mitigation of information problems, we cannot document a significant impact on the observed wealth effects. A possible explanation of this finding may be either a limited economic magnitude of these benefits or an incomplete or lagged valuation of these potential benefits by the capital markets. In contrast, we can show that the stock price reaction is significantly more positive for both target firms and the purchasing firms if the target shows symptoms of ineffective monitoring or existing agency problems. This observation strongly supports the notion that large outside blockholders can increase monitoring efficacy.
Third research question: Apart from the factors addressed above, research suggests that also the specific type of corporate relatedness (e.g. vertical vs. horizontal) may play an important role in understanding the motives and effects of corporate minority equity ownership. Accordingly, related empirical research on takeovers (i.e. acquisitions of majority control) indicates that corporate relatedness between the acquirer and the target may play an important role in understanding why some transactions create value and others do not (e.g. Haely, Palepu and Ruback (1992), Maquieira, Megginson and Nail (1998), Maksimovic and Phillips (2001), Fan and Goyal (2006)). Although the relevance of corporate relatedness
6
has yet not been empirically analyzed in the context of minority block ownership, there are numerous theoretical considerations that lead us to the hypothesis that the potential benefits of minority block ownership strongly depend on the type of corporate relatedness between the block purchaser and the target. On the one hand, minority equity ownership between vertically related firms may foster vertical integration, which in turn can reduce transaction costs that arise due to contractual inefficiencies in customersupplier relationships (e.g. Klein, Crawford and Alchian (1978), Williamson (1979), Grossman and Hart (1986), Hart and Moore (1990)). In addition, the market power of the vertically integrated firm may increase (Salinger (1988), Hart and Tirole (1990), Ordover, Saloner and Salop (1990)). On the other hand, corporate equity ownership in a horizontally related firm may give incentives to reduce competition and foster collusive behavior (Reynolds and Snapp (1986), Malueg (1992)). Considering these arguments, we want to provide empirical evidence on the relationship between corporate relatedness and the wealth effects of minority block purchases. In order to shed light on the economic implications of corporate relatedness, we examine a sample of 141 U.S. minority block purchases by non-financial corporations during the 1984 to 1997 period. Taking into account the complexity of corporate relatedness, we follow the methodology of Fan and Lang (2000) and use commodity flow data from U.S. input-output (IO) tables provided by the U.S. Bureau of Economic Analysis to construct quantitative measures of relatedness for each transaction in our sample. Using this methodology, we can show that block purchases between strategically related corporations lead to higher excess returns for targets, acquirers and the combined entity in most cases. We can also confirm this positive relationship when controlling for other factors such as R&D-intensity and relative size in cross-section. Although we can only document a relatively low statistical significance of our findings, we interpret these results as an indication for the potential benefits that stem from corporate block ownership between strategically related corporations. Our work contributes to both, financial research and corporate decision making. On the one hand, with each of the papers discussed above, we add to the growing body of
7
empirical literature on large blockholders. Our first paper (chapter 2) helps to gain a better understanding of the consequences of block purchases in a bank-based economy, taking into account both the motives of block purchasers and target characteristics such as ownership structure. The results provided in our second paper (chapter 3) help to fill a substantial research gap with regard to the determinants of wealth creation from corporate minority block purchases. Finally, the findings of our third paper (chapter 4) show that the specific type of corporate relatedness helps to further explain the wealth effects of corporate block purchases. This result is especially valuable for future studies on large (corporate) blockholders, as the role of corporate relatedness should receive increased emphasis. On the other hand, our findings may be also relevant for corporate practitioners and investment professionals engaged in the acquisition and management of corporate ownership stakes. Our theoretical considerations with regard to the potential drivers of wealth creation in the context of corporate block purchases and the related empirical evidence provided on the actual valuation consequences may serve as an initial point of reference for (corporate) investment decisions.
9
2
New Outside Blockholders, Performance, and Governance in Germany
2.1
Introduction
Many theoretical analyses argue that the involvement of large shareholders in monitoring or control activities has the potential to limit agency problems (Shleifer and Vishny (1986), Maug (1998), and Noe (2002)).1 Empirical support for this argument has been provided by a number of studies (e.g. Agrawal and Mandelker (1990), Hartzell and Starks (2003), Holderness (2003)). Results are, however, still ambiguous. Most of the studies are static in nature, examining the impact of already existing large blockholders on firm performance or monitoring related variables such as management turnover. In contrast, the body of empirical research on the dynamic effects of changes in ownership structure is considerably smaller. This strand of literature investigates acquisitions of 5% or more, that is, the emergence of new blocks. These “partial acquisitions” are typically greeted by positive abnormal returns. This reaction has been ascribed to factors such as the anticipation of an increased takeover probability and enhanced monitoring (e.g. Choi (1991), Barclay and Holderness (1991)). Recent studies further document a strong influence of block- and target-specific characteristics on the realized wealth effects (e.g. Bethel et al. (1998), Allen and Phillips (2000), Park et al. (2008)). So far, block purchases appear fairly well explored. However, the research discussed above focuses entirely on the U.S. In contrast, evidence on the role and effect of large blockholders in Europe is surprisingly scarce. We want to fill this gap by empirically examining the effect of block purchases in Germany. The German system of corporate control and governance has several unique characteristics that motivate us to specifically focus on block purchases in Germany. First, Germany in particular is characterized by a substantially higher ownership concentration than the U.S. (La Porta et al. (1999), Faccio and Lang (2002)). Becht and 1
A large shareholder (blockholder) is generally defined as an entity that owns at least 5% of a firm’s outstanding shares.
10
Boehmer (2003) document that over 80% of officially exchange listed German corporations have a blockholder controlling over 25% of the votes and that 65% are majority controlled. A considerable power can be associated with blocking minority stakes (larger than 25%), even in cases where a controlling shareholder or coalition of controlling shareholders exists (Jenkinson and Ljungqvist (2001)). Conversely, this raises the question if the capital market also values incremental monitoring by an additional large blockholder. For example, activist investors that are often associated with superior monitoring abilities (Bethel et al. (1998), Park et al. (2008)) may not be able to undergo substantial corporate restructurings in the presence of blocking minorities. Second, in the light of substantial ownership concentration, a relative active market exists in stakes. In contrast to the U.S., acquisitions of large stakes are regarded as a substitute for the traditional market of corporate control (Franks and Mayer (2001), Köke (2004)), leading to both friendly and hostile control changes (Jenkinson and Ljungqvist (2001)). Given the above considerations, we will shed light on the economic implications of block purchases in Germany by studying a new set of data with 85 block purchases with German targets between 1997 and 2006. We focus on de-facto minority block purchases between 5% and 49.9% of target voting stock to separate the wealth effects of implicit majority control transfers2 from block- and target-specific effects such as acquirers’ intention (e.g. activist investor) or target firms’ incumbent ownership structure. Our results show that block purchases in German companies lead to positive abnormal returns on average, regardless of the identity of the buyer. In addition, we find block purchases by activist investors to be accompanied with significantly higher abnormal returns than purely financial block purchases. This finding leads us to the conclusion that activist acquirers either have superior monitoring or target selection abilities. With regard to the influence of the ownership structure, we find that incumbent large shareholders have at most a marginal effect on the market reaction, despite the con2
Faccio and Stolin (2006) show that bidding firms may use partial acquisitions (acquisitions of majority control but not of 100% control) to expropriate the target firms’ minority shareholders. Accordingly, Martynova and Renneboog (2006) document that the abnormal returns to the announcements of partial acquisitions to target shareholders are substantially lower than for full acquisitions.
11
siderable power that is associated with blocking minority stakes. Finally, we show that the share price reaction is negatively correlated to target firms’ market capitalizations and valuation levels. The remainder of the paper is structured as follows: Section 2.2 discusses theoretical considerations with regard to the potential wealth effects of block purchases and the related empirical evidence. In section 2.3 we discuss the sample selection procedure, relevant block- and firm-specific data and the econometric strategy. In section 2.4 we present univariate and multivariate analyses of share price announcement effects. Section 2.5 summarizes the findings and concludes.
2.2
Literature Review
2.2.1 General Considerations with Regard to the Potential Wealth Effects of Block Purchases Considering the U.S. evidence, the majority of the early studies on the impact of block purchases on corporate performance find positive abnormal returns for the target upon the announcement of an accumulation of 5% or more of the common stock by an outside investor. In general, there are three main hypotheses that explain the source of value creation to target shareholders. The “anticipated takeover bid hypothesis” is based on the notion that block purchases (or “toehold acquisitions”) can start an inter-firm investment process, implying that the positive valuation effect is due to investors’ perception of an increased probability of a subsequent takeover. Using a sample of 337 new large block formations during the period 1978-1980, Mikkelson and Ruback (1985) examine the valuation effects of toehold acquisitions. They document that block purchases that are not followed by investment outcomes (e.g. takeovers, targeted repurchases) lead to significant negative abnormal returns for target shareholders in the post-toehold acquisition period. These results imply that the value increase at toehold acquisitions is an adjustment for the expected takeover premiums, and that the absence of takeovers makes investors revise downward the expected benefits of future takeover bids.
12
Choi (1991) and Sudarsanam (1996) confirm these findings and further identify the increased probability of external or internal control transfer as a potential source of value gains to target shareholders. Accordingly, the “monitoring hypothesis” (or “control transfer hypothesis”) states that block purchases increase the probability of control trans-fers, which is expected to increase the target firm’s management efficiency and thus future performance. Supporting the monitoring hypothesis, Holderness and Sheehan (1985) and Barclay and Holderness (1991) find that block purchases are typically followed by abnormal stock price appreciations and also increased management turnover for both purchases by “corporate raiders” and negotiated block trades, respectively. Finally, the “undervaluation hypothesis” states that block purchases are generally attempted by investors who either possess private information or superior security analysis skills, enabling them to ascertain that the target’s shares are temporarily undervalued. There is, however, no robust empirical validation of this hypothesis, given the methodological problems to identify exactly a state of temporary undervaluation (Holderness and Sheehan (1985), Shome and Singh (1995)). Since the mid 1990s, empirical evidence on the wealth effects of block purchases has focused increasingly on cross-sectional relationships between the observed share price reactions and block- or firm-specific characteristics. With regard to block-specific characteristics, especially the identity and intention of the acquirer has drawn attention. Shome and Singh (1995) document significantly higher abnormal returns to target shareholders if the acquirer of the block is either a corporation or an institutional investor. They attribute this finding to increased market power and future synergies and to efficient-monitoring. Bethel et al. (1998) use a comprehensive sample of 244 block purchases from 1980 to 1989 and introduce a more precise classification of blockholder identity. Distinguishing between activist, strategic and financial blocks, they document significant abnormal returns for target shareholders upon the announcement of block purchases by activist investors. In contrast, they only find marginally positive and slightly negative abnormal returns for financial and strategic blocks, respectively.
13
Recent studies also take into account the interaction between block- and targetspecific characteristics. For example, Akhigbe, Madura and Spencer (2004) and Park et al. (2008) document a significant impact of (control related) firm-specific characteristics. Basing their argumentation on the monitoring hypothesis, they state that the incremental benefit from an additional outside shareholder significantly depends on the target firms’ existing internal and external monitoring mechanisms such as managerial ownership or leverage.
2.2.2 Empirical Evidence on The Effects of Block Purchases in Germany Whilst there has been considerable empirical support for positive wealth effects in the U.S., evidence on the impact of block purchases in Germany is sparse. Franks and Mayer (2001) specifically focus on the wealth effect of block purchases in Germany. Analyzing a sample of 57 block purchases for the period 1988-1997, they report small positive abnormal returns of 2.3% for the [-5;5] event window. They do not provide further analyses with regard to the potential drivers of the wealth effects. The size of abnormal returns reported is substantially lower than in similar U.S. studies. For example, Bethel et al. (1998) and Park et al. (2008) report abnormal returns of up to 15.7% and 23.3%, respectively, depending on the identity of the block purchaser. Croci (2007) analyses 136 block purchases by 15 well-known corporate raiders (e.g. Guy Wyser-Pratte) in Europe between 1990 and 2001. Analyzing a small subsample of 13 German targets, he reports positive abnormal returns of 9.1% over the [-30;5] event window. Although this observation is not conclusive due to the limited sample size, we regard the higher returns reported by Croci (2007) relative to those reported by Franks and Mayer (2001) as a possible indication for the influence of cross-sectional factors such as acquirer identity on the wealth effects of block purchases. To sum up, there is no consistent empirical evidence on capital market reaction on the announcement of block purchases in Germany or other comparable European governance systems. A clarification of the value consequences of block purchases in this
14
bank-based corporate governance system will offer implications for the necessity of political/legal adjustments.
2.3
Data and Methodology
2.3.1 Sample Selection Relevant block purchases were identified using the SDC/Thomson One Banker Deals database. The initial sample comprises all 12,328 completed M&A transactions with German targets from 1997-2006. We then use 9 criteria to further specify the sample: (1) The acquired block size is 5% to 49.9% (2) The initial ownership of the block acquirer is 0% (3) The target is no financial services firm (4) The target is a public company and daily stock prices for the period t-220 to t+20 are available (5) De-facto minority acquisition (no indirect formation of majority control blocks) (6) Acquisition by outside investor (7) Identity of acquirer and specific size of acquired stake known (in % of voting stock) (8) Accounting data and consistent ownership information available (9) No material confounding news in the [-5;5] announcement window around the announcement day t0
Following Choi (1991), we define outside minority blocks as investors who own more than 5% but less than 50% of the target’s firm voting stock after the purchase (criterion (1)). As we are only interested in the formation of new outside blocks, we require the initial ownership to be 0% (criterion 2). We further exclude financial services firms (criterion (3)) to guarantee the comparability of accounting data across the sample com-
15
panies. Using criterion (4), we only include companies with sufficient return data for the estimation of the market model parameters. Applying criteria (1) to (4), we arrive at 218 transactions. As we are only interested in de-facto minority block purchases by outside investors, we apply criteria (5) and (6) to control for minority purchases that lead to de-facto control blocks (e.g. if the acquirer was holding additional stock options prior to the reported minority block purchase) and for minority block purchases of insiders or affiliated companies. The control for the formation of de-facto control blocks is particularly important, since private benefits of control and thus the potential of minority shareholder expropriation are relatively high in Germany (Nenova (2003), Dyck and Zingales (2004)). The inclusion of acquisitions of de-facto majority control would probably lead to biased results. Hence, we use the Factiva news database and ownership information from Hoppenstedt Börsenführer3 and the BaFin database on holdings of voting rights (“Stimmrechtsdatenbank”)4 to identify defacto acquisitions of majority control as well as block purchases by insiders. Applying criterion (7), we only include those deals where the identity of the acquirer and the exact size of the acquired block is known. The exact identity of the investor is important for the further cross-sectional analysis of abnormal returns. For the same reason, we are also interested in the exact size of the acquired stake by investor. Thus, we have to exclude most consortial block acquisitions, as the acquired stake size is generally reported as a total and not by acquiring party. Criterion (8) ensures that we can test hypotheses with regard to the firm-specific financial performance and ownership structure. We obtain accounting data from Thomson’s Worldscope database and ownership information for the quarter before and after the block purchase from Hoppenstedt Börsenführer, BaFin Stimmrechtsdatenbank and the ownership module of Thomson One Banker. Combined, 3
The Hoppenstedt Börsenführer is a quarterly stock guide that contains profiles and financial data for all listed German companies. 4 The German Federal Financial Supervisory Authority (BaFin) has drawn up a consolidated overview of the holdings of voting rights in German companies listed on the first segment of the German stock exchanges, given the publication requirements set out in the Securities Trading Act (WpHG). This database on holdings of voting rights is based exclusively on data gathered from the statutory publications made in supra-regional official stock exchange gazettes (e.g. Börsen-Zeitung). For the purpose of our study, the BaFin has kindly provided us with historical ownership information from this database.
16
these sources offer a fairly precise picture of the target’s ownership structure in the quarter prior to the transaction. Finally we control for material events such as M&A transactions or earnings surprises (criterion (9)) within the [-5;5] period around the block purchase at t0 in order to isolate the economic effect of the transaction. Our final sample contains 85 transactions that fulfill all of the above criteria (see Table 2-1). The sample represents 0.7% of all completed M&A transactions with a German target or 6.8% of all block purchases that fulfill criteria (1) and (2) in the SDC/ Thomson One Banker Deals database over the 10 year period from 1997 to 2006. The distribution of transactions by year and block type is documented in Table 2-2. Table 2–1: Sample Selection This table shows the total number of completed M&A transactions with German target over the 10 year period from 1997 to 2006 initially obtained from the SDC/Thomson One Banker Deals database and the number of transactions included after the application of defined selection criteria (1) to (9). Number M&A transactions with German target (1997-2006) Screened after criteria (1) and (2) Screened after criterion (3) Screened after criterion (4) Screened after criteria (5), (6) and (7) Screened after criterion (8) Screened after criterion (9)
12.328 1.252 1.134 218 144 100 85
% of all reported % of initial minortransactions ity block purchases 100.0% 10.2% 100.0% 9.2% 90.6% 1.8% 17.4% 1.2% 11.5% 0.8% 8.0% 0.7% 6.8%
Table 2–2: Distribution of Transactions by Year and Block Type This table shows the distribution of block purchases by acquirer type and year for the years 1997 to 2006. Year
Activist blocks
Financial blocks
Strategic blocks
Total
Proportion
1997
0
0
0
0
0.0%
1998
0
1
1
2
2.4%
1999
1
0
0
1
1.2%
2000
0
1
4
5
5.9%
2001
1
2
3
6
7.1%
2002
3
5
4
12
14.1%
2003
4
9
2
15
17.6%
2004
1
6
4
11
12.9%
2005
3
8
3
14
16.5%
2006
5
9
5
19
22.4%
Total
18
41
26
85
100.0%
17
2.3.2 Block Characteristics When examining the wealth effect of new block formations for target shareholders, one has to take into account that the explanation of the effect could be a distribution of both, block characteristics and target specific factors. With regard to block characteristics, several studies show that there are significant differences between the various types outside blockholders, especially among institutional investors (e.g. Gaspara, Massa and Matos (2005), Cronqvist and Fahlenbrach (2007)). Hence, the use of a formal classification scheme for the type of outside investor does not seem appropriate to capture the effect of the underlying economic intention of the individual block purchase.5 For example, a non-financial company may on the one hand build up a minority position in another non-financial company with the intention to found or to further strengthen a strategic collaboration. On the other hand, the company may also regard the investment as purely financial and not even intend any business affiliation with the target. Taking into account this differentiation, we follow Bethel et al. (1998) who distinguish between three types of outside block purchases according to their inherent intention: (1) activist blocks, (2) strategic blocks and (3) financial blocks. They define activist block purchases as those made by investors that are known for activist policies in the past and those with announced intention to influence corporate policy. Block acquisitions that are made by other companies and that are unopposed by management are considered strategic. Financial block purchases are those made by banks, pension funds, money managers, and individuals who do not act as an activist investor. Following this definition, we require for a block to be considered as an activist that the acquirer explicitly states that he will attempt to actively influence the management of the firm in order to increase firm value or to be known for activist policies in the past.6 Empirical research shows that activist investors such as hedge funds may have a significantly positive effect on wealth creation (e.g. Bethel et al. (1998), Park et al. (2008), 5
We would regard as a formal classification scheme the economic sector based classification of the German Central Bank (based on the European System of Accounts (ESA 95)). 6 In our sample, this category includes individual activist investors (e.g. Guy Wyser-Pratte) as well as institutional activist investors.
18
Brav, Jian, Partnoy and Thomas (2008)). Consequently, we expect the wealth effect of activist blocks to be higher than that of financial blocks. With regard to strategic block purchases, Chan, Kensinger, Keown and Martin (1997) show a positive market reaction to announcements of both horizontal and non-horizontal strategic alliances, even when there is no equity participation. Allen and Phillips (2000) show that block purchases that are accompanied by strategic agreements and alliances result in significantly larger excess stock returns to target shareholders when compared to block purchases by companies that are not associated with strategic intentions. Hence, we further narrow our definition and consider only those blocks as strategic where either the acquirer or the target management explicitly states that the transaction rational is strategic. In turn, the block purchases by corporations that do not have a strategic intent are classified as financial block purchases in our sample. In order to systematically classify the block purchases, we screen all news announcements around the block purchase in the Factiva database for relevant information and classify the transactions accordingly. Apart from the type of intention, the size of the acquired block is likely to affect potential monitoring activities by the partial acquirer. Blockholders owning large stakes are more likely to monitor management actions since they face a liquidity problem (Maug (1998)). Furthermore, the likelihood that the benefit of monitoring exceeds the cost increases with the size of the blockholding (Park et al. (2008)). Consequently, we include into our sample the proportion of target voting stock acquired as a measure for block size. We further identify those block purchases made with stated takeover intention, since the observed wealth effect from these transactions might partially reflect an anticipated change in majority control. Finally, we also identify those block acquisitions that are made via private placements instead of an open market transaction or a block transfer. Several studies (e.g. Hertzel and Smith (1993), Hertzel, Lemmon, Linck and Rees (2002), Barclay, Holderness and Sheehan (2007)) report a positive market reaction to most types
19
of private placement announcements.7 Hence, we control for this effect in order to correctly assess the economic impact of the corporate governance related block- and targetspecific characteristics. Table 2-3 presents descriptive statistics on block characteristics. Financial blocks represent nearly half (48.2%) of the total transactions in our sample. Activist and strategic block purchases account for 21.2% and 30.6% of all transactions, respectively. The average size (% of voting stock) of the block purchase is 16.5%. Comparing the block parameters by block type, we find that strategic blocks are on average larger than activist and financial blocks. They also inherit a takeover intention more frequently. Naturally, none of the purely financial blocks in our sample has a takeover intention. Table 2–3: Sample Description of Block Characteristics This table describes the data collected for the description of new block characteristics. The acquired block size is measured as the proportion of targets’ common stock acquired. Takeover intention is a binary variable that takes on the value of one if the block purchase is made with an expressly stated takeover intention. Private placement is binary variable that takes on the value of one if the transaction is a private placement. The percentages for acquired block size, takeover intention and private placement represent the fraction of the total number transactions by block type (i.e. by column). All blocks Acquiring block type Number of observations Percentage of total sample Acquired block size 5% - 15% 15% - 25% 25% - 50% Mean Median Stdev. Takeover intention Number of blocks Percentage of blocks Private placement Number of blocks Percentage of blocks
7
Activist blocks
Strategic blocks
Financial blocks
85 100.0%
18 21.2%
26 30.6%
41 48.2%
51.8% 18.8% 29.4%
61.1% 11.1% 27.8%
23.1% 19.2% 57.7%
65.9% 22.0% 12.2%
16.5% 14.0% 10.3%
15.1% 10.0% 10.7%
23.0% 25.0% 9.4%
13.0% 10.0% 8.7%
6 7.1%
2 11.1%
4 15.4%
0 0.0%
11 12.9%
3 16.7%
3 11.5%
5 12.2%
There are three general hypotheses that explain the value effect of private placements: monitoring, value certification and management entrenchment. For a detailed theoretical description and empirical analysis see Barclay, Holderness and Sheehan (2007).
20
2.3.3 Target Characteristics Apart from block-specific characteristics, we are interested in the ownership structure of the target company. The presence of large shareholders is generally regarded as beneficial, since their monitoring or control activities have the potential to limit agency problems. Conversely, this implies that if the target company already has one or more large blockholders prior to the block purchase, the incremental monitoring-related benefit from an additional large blockholder is likely to diminish. Thus, we include three proxy variables into our sample to control for target ownership structure. We first count the number of ownership blocks holding at least 5% of voting stock in the quarter prior to the block purchase. We then measure ownership concentration as the total proportion of voting stock held by these large blocks. We also identify those targets that have at least one controlling shareholder holding more than 25% of voting rights.8 In this case, at least one incumbent shareholder holds a blocking minority (25-50%) or even majority control (>50%). This measure is highly relevant in the context of the German corporate governance system, given the considerable power associated with blocking minority stakes (Jenkinson and Ljungqvist (2001)). Hence, we expect lower incremental benefits from a new large blockholder in the presence of incumbent large blockholders and thus a weaker capital market reaction. Second, we include additional financial information about the target companies that have a potential effect on the market reaction. We distinguish between factors affecting potential benefits through incremental monitoring and other factors. Considering factors that relate to the monitoring hypothesis, Helwege, Pirinsky and Stulz (2007) assume that the level of information asymmetries decreases with firm size. They argue that larger firms are more frequently monitored by analysts, institutional investors and regulators. We therefore measure the target’s market capitalization prior to the block purchase (t-21) as a proxy for firm size. Debt can be regarded as an external corporate governance device, since it can reduce the agency cost between management and owners (Jensen (1986)). The higher the target’s level of debt prior to the block purchase, the lower we expect the 8
Although German law basically allows an investor owning more than 50% of all shares to appoint management, owning more than 25% of the voting stock grants the right to veto decisions.
21
incremental monitoring benefit from a new blockholder. Therefore, we include the debt ratio into our sample to proxy the extent of the monitoring effect of debt. Unrelated to the monitoring hypothesis, stake purchases might also be perceived as a signal that the target company is undervalued (Banerjee, Leleux and Vermaelen (1997)). We therefore include the market-to-book ratio to proxy target firms’ valuation levels. Table 2-4 summarizes target firm characteristics of the sample. Table 2–4: Sample Description of Target Characteristics This table describes the data collected for the description of target characteristics. Block concentration is measured as the sum of voting stock held by all blocks >5%. The number of ownership blocks represents the number of all ownership blocks >5% present in the quarter prior to the transaction. Presence of controlling shareholder is a binary variable that takes on the value of 1 if the target firm has an incumbent large shareholder holding at least 25% of voting stock in the quarter prior to the transaction. Market-tobook ratio is defined as the market value of common equity to the book value of common equity at the end of the estimation period (t-21). Debt ratio is defined as the as the book value of debt (Worldscope item 03255) divided by the firm value (Worldscope item 03255 + market value of equity at t-21). Market value represents the market value of targets’ common equity at t-21.
Ownership characteristics Block concentration (>5% blocks) Mean Median Number of large blocks Mean Median Presence of controlling shareholder Number of blocks Percentage of blocks by type Financial characteristics Market-to-book ratio Mean Median Debt ratio Mean Median Market value [€ Mio.] Mean Median
All targets
Activist block targets
Strategic block Financial block targets targets
44.4% 44.3%
42.3% 42.1%
42.6% 38.6%
46.5% 50.5%
2.3 2.0
2.7 2.5
2.5 2.0
2.1 2.0
50 58.8%
9 50.0%
15 57.7%
26 63.4%
5.3 1.5
1.5 1.3
2.5 1.8
8.8 1.5
36.1% 33.3%
31.5% 23.8%
40.5% 44.6%
35.4% 33.3%
797.8 86.7
332.8 212.2
118.8 52.5
1432.6 112.7
22
2.3.4 Method of Measuring Abnormal Returns To assess the value implication of the block purchase on the remaining shareholders, we use standard event study methodology. Expected returns are generated from the market model parameters, estimated with daily returns from 220 days before the block purchase announcement (t-220) to 21 day before the announcement (t-21). Adjusted prices (taking into account dividend payments and relevant changes to the capital structure) are used to calculate stock returns. The CDAX9 is used as a proxy for the market returns. Abnormal returns (ARs) are calculated as the difference between actual returns and estimated returns from the market model in the event window. Following the suggestion by Harrington and Shrider (2007), we use the test statistic of Boehmer, Musumeci and Poulsen (1991) to test the significance of cumulated abnormal returns.
2.4
Results
2.4.1 Announcement Effects Table 2-5 documents the average cumulative abnormal returns and test statistics for the total sample (panel A) and for the three subsamples by block type (panel B, C and D). The results from panel A suggest that purchases of minority blocks lead to a statistically significant positive market reaction. The average abnormal return amounts to 7.97% (z = 7.56) in the event window [-1,1]. Panel B and C show that both activist and strategic block purchases lead to significant positive announcement effects. Average cumulative abnormal returns in the event window [-1;1] amount to 11.75% (z = 6.55) for activist blocks. Although the sample size for activist blocks lets us use this finding with some caution, we evaluate this finding as a preliminary confirmation of the results of Bethel et al. (1998), Park et al. (2008) and Croci (2007). We also find strategic block purchases to create significant value to target shareholders. The average cumulative abnormal return of transactions with strategic intend is 11.32% (z = 5.43) in the event 9
The CDAX (or composite DAX) is a German all-share index introduced in 1993 that covers all German shares that are admitted to the Prime Standard and General Standard segments. Therefore, the index reflects the performance of the overall German equity market and is thus well suited for the purpose of this study.
23
window [-1;1]. Panel D shows that the market also reacts positively to financial block purchases. However, the wealth effect is significantly lower than for activist and strategic blocks. Table 2–5: Excess Returns by Acquiring Block Type This table shows the average cumulative abnormal returns for the total sample 85 minority block acquisitions and for the 3 subsamples by acquirer category. Statistical significance at the 1%, 5%, and 10% level is denoted with ***, **, *. The statistical significance is tested using the test-statistic of Boehmer, Musumeci and Poulsen (1991) (z-statistics) and the standard t-test (t-statistics). Event time period [-20;20] [-10;10] [-5;5] [-2;2] [-1;1] [-1;0] Event time period [-20;20] [-10;10] [-5;5] [-2;2] [-1;1] [-1;0]
Panel A: Total sample (N=85) CAAR t-statistics z-statistics
Panel B: Activist blocks (N=18) CAAR t-statistics z-statistics
10.71%*** 3.54*** 4.42*** 9.38%*** 5.59*** 5.74*** 8.71%*** 6.47*** 6.49*** 8.15%*** 6.61*** 6.80*** 7.97%*** 6.52*** 7.56*** 5.33%*** 5.54*** 5.97*** Panel C: Strategic blocks (N=26) CAAR t-statistics z-statistics 13.06%*** 2.39*** 2.75*** 15.70%*** 4.45*** 5.21*** 12.66%*** 5.12*** 5.79*** 12.28%*** 4.34*** 4.89*** 11.32%*** 4.41*** 5.43*** 8.15%*** 3.69*** 3.99***
12.06%*** 2.73*** 3.59*** 11.50%*** 5.22*** 5.81*** 11.92%*** 5.94*** 6.69*** 11.18%*** 7.33*** 6.23*** 11.75%*** 6.16*** 6.55*** 6.15%*** 3.39*** 3.18*** Panel D: Financial blocks (N=41) CAAR t-statistics z-statistics 8.64%*** 1.75*** 1.94*** 4.43%*** 1.95*** 1.50*** 4.79%*** 2.40*** 1.94*** 4.20%*** 2.79*** 2.63*** 4.18%*** 2.63*** 3.01*** 3.17%*** 2.86*** 3.33***
Figure 2-1 illustrates a dynamic view on the development of the cumulative abnormal returns in the [-10;10] event window. The results show that there is a run-up in stock prices before the event. There are two possible explanations for this observation. First, a large amount of the block purchases is carried out through open market purchases, and it can take months to accumulate a stake for which an official notification is required. As a consequence, the market may already react to increased buy-side pressure before an official announcement of a new block formation.
24
Second, there is a potential lag between the actual block purchase and the mandatory public announcement.10 Hence, it cannot be ruled out that information on the transaction is already known to individual market participants prior to the official announcement date. Figure 2–1: Cumulative Average Abnormal Target Returns This figure illustrates the development of cumulative average abnormal returns from day t-10 to t+10.
Cumulative average abnormal return
20% 15%
Activist blocks Strategic blocks Financial blocks
10% 5% 0% -5% -10 -9 -8 -7 -6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6 7 8 9 10 Days relative to announcement
Summing up, our approach to examine minority block purchases in Germany reveals a value creation to targets’ shareholders for all minority block types. At the first glance, the univariate analysis also seems to confirm our initial hypothesis that activist blocks and strategic blocks hold more potential for value creation than financial blocks. However, the sample statistics presented in tables 2-3 and 2-4 suggest systematic differences in block and target company characteristics. In order to control for these factors, we conduct multivariate analyses in the following section.
10
§21 AktG (corporate law) requires that target management must be notified immediately if another corporation’s engagement exceeds 25% or 50% of the target’s voting rights. In addition, the Securities Trading Acct tightened the disclosure standards so that all investors have to notify the Financial Supervisory Authority upon arriving at the 5, 10, 25, 50, or 75 percent threshold, be it from above or below. Finally, the Transparency Directive Implementation act which came into force on January 20th 2007 added the thresholds of 15, 20, and 30 percent; this law also introduced a 3 percent threshold level to prevent “creep-ins”. Since we do not want to address the complex process of voting rights disclosure in this context, we refer to Becht and Böhmer (2003) who discuss several transparency issues that arise from the current legal provisions covering the disclosure of control rights.
25
2.4.2 Determinants of Abnormal Returns We design four regression models to determine the drivers of the observed wealth effects to target shareholders. The abnormal return in the period [-1;1] serves as dependent variable.11 Table 2-6 summarizes design and results for the four regression models. All test statistics are computed using White’s heteroskedasticity-consistent covariance matrix (White (1980)). The variables included into our model are structured following our initial hypotheses on the factors that potentially explain the observed wealth effect. We include five variables to control for block-specific characteristics: “activist block”, “strategic block”, “takeover intention” and “private placement” are each dummy variables taking on the value 1 if the attribute is true. “Block size” is a variable that measures the size of the new block as the proportion of voting stock acquired. We also proxy target company characteristics via a predefined set of variables and distinguish between target ownership characteristics and target financial characteristics. A total of three variables is used to describe the ownership structure of the target prior to the acquisition. “Concentration of existing blocks” equals the combined share of voting rights of all blocks bigger than or equal to 5%. “Number of blocks” represents the number of all blocks larger than 5%. The dummy variable “presence of controlling shareholder” takes on the value of 1 if the target company has a controlling blockholder (>25% voting rights) in the quarter prior to the block purchase. With regard to the target’s financial characteristics, we include three variables (“logarithm of market value”, “debt ratio”, and “marketto-book ratio”). In order to shed light on the drivers of the wealth effect to target shareholders, we first design a model that does not control for the intentions of the block acquirer. Hence, we omit the dummy variables “activist block”, “strategic block” and “takeover intention” from the estimation model 1. We then include these variables again in models 2 to 4, where we explicitly control for the type of blockholder and takeover intention. Models 2 to 4 only differ in the variables used to control for the targets’ ownership structure prior
11
Using the abnormal return over the [-10;10] event window as the dependent variable leads to the same qualitative results.
26
to the acquisition, since the underlying economic hypotheses for each variable differ from each other. We refrain from jointly using the variables “concentration of existing blocks”, “number of existing blocks” and “presence of controlling shareholder” in one model to avoid a multicollinearity problem. Comparing model 1 to the other models, we see that the inclusion of block intention significantly influences the regression results. Table 2–6: Determinants of Cumulative Average Abnormal Target Returns This table shows estimation results for OLS-regression models with the cumulative abnormal return for the [-1;1] interval as the dependent variable. Activist block, strategic block, takeover intention, private placement and controlling shareholder are binary variables that take on the value 1 if the attribute is observable in the given transaction. t-values are in brackets. Statistical significance at the 1%, 5%, and the 10% level is denoted with ***, **, *. All test statistics are computed using a heteroskedasticity-consistent covariance matrix (White (1980)). Model 1 Intercept
0.112)*** (1.673)* **
0.003)*** (1.971)***
-0.034)*** (-1.001)***
0.059)*** (2.18)** * 0.016)*** (0.345)*** 0.002)*** (0.961)*** 0.042)*** (0.984)*** -0.035)*** (-1.026)***
0.067)*** (2.513)** * 0.024)*** (0.566)*** 0.002)*** (0.881)*** 0.053)*** (1.155)*** -0.027)*** (-0.796)***
0.059)*** (2.166)** * 0.018)*** (0.416)*** 0.002)*** (0.94)*** 0.037)*** (0.802)*** -0.029)*** (-0.858)***
-0.001)*** (-1.986)***
-0.001)*** (-1.572)***
Takeover intention
Target ownership characteristics Concentration of existing block Number of existing blocks
-0.012)*** (-1.795)***
Presence of controlling shareholder Target financial characteristics Logarithm of market value Debt ratio Market-to-book ratio N Adj. R2 F-statistic
Model 4
0.129)*** (2.265)** *
Strategic block
Private placement
Model 3
0.122)*** (2.077)***
Block characteristics Activist block
Block size
Model 2
0.134)*** (2.189)***
-0.023)*** (-0.994)***
-0.016)*** (-1.932)***
-0.016)*** (-1.914)***
-0.017)*** (-1.969)***
-0.017)*** (-1.885)***
0.031)*** (0.653)*** -0.001)*** (-6.622)*** 85 14.9% 3.454***
0.042)*** (0.927)*** -0.001)*** (-3.155)*** 85 17.5% 2.974***
0.024)*** (0.595)*** -0.001)*** (-3.057)*** 85 18.0% 3.056***
0.039)*** (0.902)*** -0.001)*** (-3.042)*** 85 16.4% 2.831***
27
The coefficients of the variables “block size” and “concentration of existing blocks” still have the expected sign, but are not statistically significant anymore. In contrast, the coefficient of the activist dummy is positive and highly significant for models 2, 3 and 4, suggesting that activist blocks have a greater monitoring potential than purely financial blocks. This result supports the findings of Bethel et al. (1998) and Park et al. (2008). However, it may also be the outcome of systematic differences in target selection. Cronqvist and Fahlenbrach (2007) provide evidence that block purchases by activists can be rather associated with influence on corporate policy and firm performance than with a selection explanation. In line with the U.S. evidence we interpret our findings in a way that the positive abnormal returns around activist block purchases in our sample stem from positive capital market expectations with regard to beneficial monitoring activities. Inconsistent with the univariate results, the regression models show that strategic blocks do not have a significant positive effect when abnormal returns are controlled for block and target characteristics. Block size and takeover intention both have a positive coefficient as expected, but are not statistically significant. Consistent with model 1, the dummy variable for private placement has a negative and insignificant coefficient. Considering the ownership structure of the target prior to the block purchase, we find negative coefficients for all proxy variables as initially hypothesized. However, only the “number of existing blocks” variable in model 2 shows a statistical significance. If the capital market valued either ownership concentration or the presence of large blockholders, we would expect an immediate market reaction to an announcement of any substantial change affecting these factors. This result indirectly supports recent evidence by Beiner, Drobetz, Schmid and Zimmermann (2006) who show that neither the presence of a controlling shareholder nor large blockholders have a significant valuation impact. The finding is also consistent with Dherment-Ferere et al. (2004) who report that large shareholders do not play a significant role in disciplining underperforming management. Regarding targets’ financial characteristics, the debt ratio has no significant impact on the market reaction. Although we would have expected a negative correlation between leverage according our initial hypothesis, this result supports Dherment-Ferere et al.
28
(2004) who cannot find robust evidence of creditor monitoring in German companies with high leverage. The coefficients for firm size and market-to-book ratio are negative and significant also when we control for block intention. With regard to firm size, this result suggests that the incremental benefit from new outside blockholders in large firms is likely to diminish, given the increased extend of external monitoring by analysts, regulatory authorities and other parties. Our finding that abnormal returns to target firms’ shareholders tend to be higher for companies with low valuation levels (reflected by a low market-to-book ratio) may be subject to two different effects. On the one hand, the capital market might perceive a block purchase in a target with a low market-to-book ratio as an undervaluation signal. On the other hand, one can interpret this result as an indication for decreasing incremental monitoring benefits from new large blockholders in the presence of high growth opportunities. As the discretion on the part of management may be higher in high growth companies (Helwege et al. (2007)), external monitoring by large outside blockholders might become inefficient.
2.5
Conclusion and Outlook
In this study we investigate share price reactions to the formation of new minority outside blocks for a sample of 85 German block purchases between 1997 and 2006. In the first part, we document a significant wealth creation for target companies, regardless of the economic intention of the acquirer. Most importantly, this finding is consistent with the U.S. evidence, despite the structural differences initially discussed. Given the low takeover activity in Germany, we attribute the observed wealth creation to either the monitoring or the undervaluation hypothesis. We then analyze whether the market reacts differently to block purchases based on block and target characteristics. We can only confirm a significant positive relationship between activist block purchases and abnormal returns when we jointly control for other block- and target-specific characteristics. However, we cannot clearly distinguish if the driver of this effect is the market’s perception of a superior monitoring ability by activist investors or rather superior stock picking skills.
29
Considering target characteristics, our results also suggest that existing target ownership structure has at most a marginal impact on the amplitude of the market reaction. We interpret this finding as indirect confirmation of recent studies that do not find a significant relation between ownership concentration and large blockholders on the one hand and firm value (Beiner et al. (2006)) or effective monitoring (Dherment-Ferere et al. (2004)) on the other hand. Validations of this interpretation in further detail should be a fruitful topic for future research.
31
3
Equity Ownership, Agency Problems and Shareholder Wealth: Understanding the Unique Role of Corporate Block Owners
3.1
Introduction
Although nonfinancial corporations frequently purchase and own equity blocks in other nonfinancial corporations in both Europe (Dherment-Ferere et al. (2004), Kirchmaier and Grant (2005)) and the U.S. (Cronqvist and Fahlenbrach (2007)), remarkably little is known about the specific motives and consequences of corporate block purchases. Relative to other types of blockholders such as banks, individual investors or pension funds, corporations have several unique characteristics that give reason to a specific analysis of the motives and consequences of corporate block purchases. Research suggests that corporate equity ownership might help to align incentives and to mitigate information problems in formal product market relationships such as strategic alliances and joint ventures (Allen and Phillips (2000)) or in supplier-customer relationships (Fee et al. (2006)). A prominent practical example of the relevance of these theoretical considerations is Daimler’s acquisition of a blocking minority stake in Tognum, a supplier of off-highway diesel engines, for €585 million in 2008. Just three years earlier, Daimler Chrysler had sold its entire interest in the company to Swedish financial investor EQT. By (re)purchasing an equity block, Daimler intended to protect its longterm supply relations with Tognum.12 Additionally, corporate block purchases may also help to mitigate information problems with regard to the specific investment opportunities of the target firms to the capital market. Allen and Phillips (2000) show that corporate blockholders are able to monitor a target firm’s management more effectively than other types of blockholders. Reynolds and Snapp (1986) argue that partial ownership agreements can reduce product market competition and thus increase the probability of monopoly rents, especially in Europe (Milanesi and Winterstein (2002), Ezrachi and Gilo (2006)).
12
See Financial Times Germany, “Daimler sichert Zugriff auf Tognum ab”, May 2, 2008, p. 10.
32
Despite these potential theoretical benefits of corporate equity ownership, especially with regard to the mitigation of potential agency problems, empirical evidence is sparse. Most empirical studies of block equity purchases focus on the general role and effect of block equity positions established in anticipation of a takeover (“toeholds”) or private equity sales.13 The majority of the block purchases in these studies are made by institutional investors and not corporations. Although several studies of block equity purchases further differentiate between different types of blockholders,14 they do not provide further insights into the specific drivers of value creation in corporate block equity purchases. The only study that explicitly takes into account the specific benefits of corporate block ownership is that of Allen and Phillips (2000). They find the largest significant increases in targets’ stock prices, investment, and operating profitability when ownership is combined with joint ventures, alliances and other product market relationships between target and purchasing firms, especially in industries with a high R&D intensity. However, they do not account for several other factors that may affect value creation from corporate block purchases such as target financial characteristics and transaction characteristics. Given the limitations of previous research, we will generate further insights into the economic implications of corporate block purchases by studying a sample of 113 European minority block equity purchases15 by nonfinancial corporations during the 1993 to 2006 period. Our main objective is to gain a better understanding of the factors that explain the differential in wealth creation from corporate block purchases. This is the first study that explicitly analyses corporate block equity purchases for a sample of European corporations.16 Hence, this paper adds to the growing body of literature that explores the market for partial corporate control and the effects of blockholders in Europe.
13
Studies that analyze these events include Mikkelson and Ruback (1985), Wruck (1989), Choi (1991), Hertzel and Smith (1993) and Ferguson (1994). 14 For example, Bethel, Liebeskind and Opler (1998) and Park, Selvili and Song (2007) distinguish between “activist”, “strategic” and “financial” purchasers. 15 Following Choi (1991), we consider acquisitions of 5% - 49.9% of outstanding shares as minority block equity purchases. 16 Croci (2007) studies block equity purchases by corporate raiders in Europe between 1990 and 2001.
33
Our approach to this study is as follows. We first give a brief overview over the theoretical benefits of corporate block ownership and the related empirical evidence. We further identify several factors that can potentially explain the observed wealth effects of the transactions. We then provide descriptive statistics on our sample and document specific transaction characteristics. Next, we examine the cumulative abnormal returns to target firms, acquiring firms, and the combined returns of targets and acquirers around the announcement of block purchases. We conduct both univariate and multivariate analyses to identify the determinants of the wealth effect for target firms and purchasing firms. More specifically, we examine whether the cross-sectional differences in excess returns can be explained based on product market relationship related characteristics, industry characteristics, transaction characteristics or target financial characteristics. Our most important results can be summarized as follows. The stock price of target firms, on average, reacts significantly positive at the announcement of corporate block purchases, indicating a significant wealth creation for target shareholders. Moreover, we find that the combined value-weighted returns are positive and significant, which indicates that corporate equity ownership between corporations is value creating in total. We also can exclude systematic wealth transfers from the purchasing firms to the target firms and vice versa. Factors that relate to the specific hypotheses regarding the potential benefits of corporate block purchases such as the alignment of incentives or the mitigation of information problems in business relationships do not seem to have a significant impact on the observed wealth effects. This may be either due to the limited economic magnitude of these benefits or to an incomplete or lagged valuation of these potential benefits by the capital markets. The stock price reaction is significantly more positive for both target firms and purchasing firms if the target shows symptoms of ineffective monitoring or existing agency problems. Finally, we also show that diversifying transactions where the acquirer purchases a block in a target firm that operates in an unrelated industry lead to significantly lower abnormal returns to acquiring firms than non-diversifying transactions.
34
The remainder of the paper is organized as follows. In section 3.2 we discuss theoretical causes and consequences of corporate block purchases and review the related empirical evidence. Section 3.3 describes the sample selection process and our research methodology. In addition, cross-sectional factors that may help to explain the potential wealth effects are identified and descriptive statistics are presented. Section 3.4 contains our empirical results and section 3.5 concludes.
3.2
Literature Review
3.2.1 Theoretical Benefits of Corporate Block Equity Ownership In this section, we discuss the potential benefits of minority equity ownership between firms in more detail. We identify several potential advantages that are unique to corporate block equity ownership and do not apply to other blockholders such as institutional investors. Firstly, a specific advantage of corporate block ownership lies in the potential alignment of incentives and the mitigation of information problems in corporate business relationships. Based on transaction-costs economics theory, Williamson (1979) suggests that equity ownership can lower the costs of monitoring agreements between firms. Allen and Philipps (2000) further pursue this notion in a detailed review of research in both industrial organization and finance and show that block ownership by corporations may be specifically useful in aligning the incentives of firms involved in business agreements, alliances or joint ventures. In this context, contracting and monitoring costs between corporations might be reduced in the presence of an equity ownership position which increases the incentives of the parties involved to invest in relationship-specific assets or product market relationships and to foreclose opportunistic behavior. Corporate equity ownership may also alleviate contractual and financing difficulties in customer-supplier relationships. The companies involved may have an incentive to partially internalize the effects of their behavior on a business partner in the presence of partial equity ownership (Mathews (2006)). Accordingly, Fee et al. (2006)) find that equity stakes are much more common in supplier-customer relationships where the sup-
35
plier is a R&D intensive firm facing a high degree of information asymmetry. This is consistent with an incomplete contracts motivation for the use of equity in inter-company relationships. More generally, Stuart, Ha and Hybels (1999) consider equity ownership as a signal for higher commitment and an additional level of confidence in a corporate partnership, which should benefit the companies regardless of the type of the business relationship. Secondly, block purchases by an outside corporation may also help to mitigate information problems regarding the specific investment opportunities of target firms and validate the target’s investment opportunities to the capital markets (Allen and Philipps (2000)). Although block purchases are frequently perceived as a signal that the target company may be undervalued (Banerjee et al. (1997)), block purchases by corporate acquirers in particular may have a strong signaling effect concerning the future prospects of the target company, given the hypothesized superior industry knowledge and operational expertise of corporate purchasers relative to other blockholders such as financial institutions (Kang and Kim (2007)). Thirdly, corporate blockholders may also be able to more effectively monitor and influence management than other types of shareholders, especially when they operate in a similar business (Spencer, Akhigbe and Madura (1998)). Accordingly, Cronqvist and Fahlenbrach (2007) find that corporate blockholders play a role for most corporate policies, whilst institutional investors such as banks, trusts and money managers do not have a significant effect. Finally, a potential reduction of product market competition may motivate corporations to establish minority equity positions. Reynolds and Snapp (1986) give a theoretical prove that if the firms involved are actual or potential competitors, partial ownership agreements may reduce competition. Although Malueg (1992) shows that the reverse may occur under certain conditions, he also confirms that in many cases, increases in cross-ownership lead to less competition and an increased likelihood of collusion. These economic considerations with regard to partial equity ownership are increasingly discussed in legal research. Several studies (e.g. Milanesi and Winterstein (2002), Russo
36
(2006), Ezrachi and Gilo (2006)) argue that minority investments among competitors can enable firms unilaterally to raise prices. Moreover, tacit or explicit collusion between competing undertakings may be facilitated. As this behavior is not regulated efficiently under the EU competition law,17 it should further increase the incentive for collusive behavior. As a consequence, the firms involved should benefit from a higher likelihood of increased monopoly rents, especially when considering the current regulation under EU competition law.
3.2.2 Empirical Evidence on the Wealth Effects of Corporate Minority Block Purchases Despite the specific characteristics and potential benefits of corporate equity ownership discussed above, empirical evidence on the wealth effect of corporate block purchases and the related determinants is fragmented and largely incomplete. Most empirical studies of block equity purchases focus on the general role and effect of “toeholds”, i.e. block equity positions established in anticipation of a takeover (Mikkelson and Ruback (1985), Choi (1991), Ferguson (1994)), or private equity sales (Hertzel and Smith (1993), Wruck (1989)). However, most of the block purchases in these studies are made by institutional investors and not corporations. Moreover, these examinations do not differentiate between corporations and other types of acquirers. Only a few studies take into account potential differences in the type of investor by explicitly differentiating between corporate purchasers and other types of acquirers. For example, Shome and Singh (1995) analyze 96 block purchases made between 1984 and 1986. They report significantly higher abnormal returns when the block purchaser is a corporation compared to block purchases by individuals. They attribute this finding to the potential for market power and future synergy gains. However, the limited number of 17
Although the European competition law provides a specific regulatory regime in case of minority shareholdings leading to collusion or de-facto control, cases that fall short of collusion or control may not be covered comprehensively as the existing regulatory instruments do not address the full range of potential anticompetitive effects of minority investments. See Ezrachi and Gilo (2006), Russo (2006) and Milanesi and Winterstein (2001) for a detailed discussion of the anti-competitive concerns with respect to minority shareholdings in the EU.
37
15 corporate block purchases included in their sample draws into question the reliability of this evidence. Bethel et al. (1998) do not find significant positive abnormal returns upon the announcement of new outside corporate blockholders. Analyzing a sample of 244 outside block formations from 1980 to 1989, they also report that firms targeted by either financial or corporate buyers do not undergo substantial operational changes and experience smaller ex-post improvements in profitability, in contrast to firms that experience activist block share purchases. This empirical finding questions the notion of superior monitoring abilities by corporate blockholders. However, the robustness of this evidence again seems questionable in the light of only 31 evaluated corporate block purchases. Contrarily, based on a sample of 264 purchases, Park et al. (2008) report that block purchases by corporate acquirers who expressly state a strategic alliance as the principal consideration of the deal generate significantly higher abnormal returns than financial institutions and regular corporate block purchasers. The only study that specifically examines the role of corporate blockholders is that of Allen and Phillips (2000) who analyze a comprehensive sample of 402 U.S. corporate block purchases from 1980 to 1991. They report significant positive announcement returns to target firms, especially in cases where product market relations such as strategic alliances or joint ventures are formed between target firms and corporate blockholders. Abnormal returns to purchasers and the combined abnormal returns of targets and purchasers are not statistically significant. Although corporate equity ownership combined with product market relationships in R&D-intensive industries lead to improvements in operating performance of target firms, insights into other potential determinants of the observed wealth effects still remain unaddressed. To sum up, little is known about the wealth effect of corporate block ownership. We fill this gap and provide further insight into the extent and the determinants of wealth gains from equity block purchases by corporations.
38
3.3
Data and Methodology
3.3.1 Corporate Block Purchases We analyze a sample of minority block purchases by European nonfinancial companies. Relevant transactions were identified using the SDC database. The initial sample comprises all 4,573 completed M&A transactions with public European targets and acquirers for the 1993 to 2006 period. We then use eight criteria to further specify the sample: (1) The acquired block size is 5% to 49.9% and the acquirer does not have a prior ownership position (2) The acquirer is an outside blockholder (3) Neither the acquirer nor the target is a financial services firm (SIC codes 60006999) (4) Acquirer and target are both public companies and daily stock prices for the period t-220 to t+20 are available in the Datastream database (5) Both acquirer and target have a market value of at least $10m (6) Accounting data are available in the Worldscope database (7) Details of the block purchase reported in the SDC database are confirmed by searching the Factiva database (8) Announcement of the block purchase does not take place in the middle of a tender offer process, subsequent to an announced takeover bid or with an expressly stated takeover intention Following Choi (1991), we define outside minority blocks as investors who own more than 5% but less than 50% of the target’s firm voting stock after the purchase (criterion (1)). By applying criterion (2), we exclude internal transactions that do not involve outside blockholders (e.g. targeted repurchases) from the sample. As we are only interested in corporate block purchases, we exclude all transactions that involve financial services firms (SIC codes 6000-6999) (criterion (3)). We only include transactions where sufficient return data is available for the estimation of abnormal returns for both target
39
and acquiring firms via the market model (criterion (4)). In order to assure that there is sufficient public information on the block trade and reasonable trading volume, we further restrict our sample to companies that have a market value of more than $10m (criterion (5)). Criterion (6) ensures that we can test the specific hypotheses with regard to target firms’ financial performance. Applying criteria (1) to (6), we derive 234 transactions. In order to maximize the reliability of our sample, we confirm each of the remaining transactions reported in the SDC database by screening news articles from the Factiva database for related information. We can find information on 224 transactions (criterion (7)). Next, we further exclude those transactions that took place in the middle of a tender offer process, subsequent to a full takeover bid or with a stated takeover intention (criterion (8)),18 since the wealth effect from these transactions is likely to partially reflect an anticipated change in majority control and not the wealth effects stemming from the specific benefits of minority corporate block ownership. Finally, in order to further enhance the quality and reliability of our sample, we use the information obtained from the Factiva database to control for potential misspecifications in the SDC database. In total, we have to manually exclude 72 transactions due to misspecifications in the SDC database or additional information obtained from the Factiva database. 19 Table 3-1 gives a detailed overview over the actual sample selection process.
18
We use information obtained from the Factiva database to identify the transactions relating to criterion (8). 19 The 72 manual exclusions were made due to the following reasons: Existence of prior ownership position not reported in SDC database (44), block purchase leading to de-facto majority control (19), transactions within one corporate entity (5), no corporate buyer (3), no equity purchase (1).
40
Table 3–1: Sample Selection This table shows the total number of completed European M&A transactions with public European target and acquirer over the period from 1993 to 2006 initially obtained from the SDC/Thomson One Banker Deals database and the number of transactions included after the application of defined selection criteria (1) to (8) and after other exclusions. The sample selection criteria are: (1) The acquired block size is 5% to 49.9% and the acquirer does not have any prior ownership; (2) the acquirer is an outside blockholder; (3) neither the acquirer nor the target is a financial services firm (SIC codes 6000-6999); (4) acquirer and target are public companies and daily stock prices for the period t-220 to t+20 are available in the Datastream database; (5) both acquirer and target have a market value of at least $10m; (6) accounting data are available in the Worldscope database; (7) details of the block purchase reported in SDC database are confirmed by searching the Factiva database; (8) announcement of the block purchase does not take place in the middle of a tender offer process, subsequent to an announced takeover bid or with an expressly stated takeover intention. Other exclusions include non-equity purchases, transactions with non-corporate buyer, transactions where the acquirer has a direct or indirect ownership position, transactions that lead to de-facto majority control, and transactions that were realized within one corporate entity. % of all reported transactions
Number
% of initial minority block purchases
European M&A transactions (1993-2006)
4,573
100.0%
Screened after criterion (1)
1,116
24.4%
-
Screened after criterion (2)
881
19.3%
100.0%
Screened after criterion (3)
458
10.0%
52.0%
Screened after criterion (4)
263
5.8%
29.9%
Screened after criterion (5)
247
5.4%
28.0%
Screened after criterion (6)
234
5.1%
26.6%
Screened after criterion (7)
224
4.9%
25.4%
Screened after criterion (8)
185
4.0%
21.0%
After other exclusions
113
2.5%
12.8%
Our final sample contains 113 transactions that fulfill all of the above criteria. The sample represents 2.5% of all completed European M&A transactions or 12.8% of all block purchases that fulfill criteria (1) and (2) in the SDC database over the period from 1993 to 2006. Tables 3-2, 3-3 and 3-4 provide information on the sample composition. The annual frequency of corporate block purchases is reported in Table 3-2. With regard to the geographic scope of the transactions, Table 3-3 shows that the largest proportion of target and acquiring firms is from France, Sweden, Germany, the UK and Spain. As shown in Table 3-4, communications, utilities & waste management and services are the industries that constitute the largest number of both target and acquiring firms.
41
Table 3–2: Distribution of Block Purchases by Year This table shows the distribution of the corporate block purchases by year. Year
Number of transactions
Percentage
Cumulative percentage
1993
3
2.7%
1994
3
2.7%
2.7% 5.3%
1995
6
5.3%
10.6%
1996
10
8.8%
19.5%
1997
6
5.3%
24.8%
1998
7
6.2%
31.0%
1999
15
13.3%
44.2%
2000
12
10.6%
54.9%
2001
12
10.6%
65.5%
2002
8
7.1%
72.6%
2003
6
5.3%
77.9%
2004
10
8.8%
86.7%
2005
5
4.4%
91.2%
2006
10
8.8%
100.0%
Total
113
100.0%
-
Table 3–3: Sample Transactions by Home Country of Target and Acquirer
Acquirer home country
This table shows the geographical distribution of the 113 sample transactions. AT: Austria, BE: Belgium, CH: Switzerland, DE: Germany, DK: Denmark, ES: Spain, FI: Finland, FR: France, GB: United Kingdom, GR: Greece, IE: Ireland, IT: Italy, NL: Netherlands, NO: Norway, PL: Portugal, SE: Sweden.
AT BE CH DE DK ES FI FR GB GR IE IT NL NO PL SE
AT 1
BE
1
CH 1
DE
2 1
3 7
DK
Target home country FI FR GB GR 1 1 2
ES
1
IT
NL
NO
PL
SE
1 2
2
1 1
11
1 2
1
1
3 2
1 12 1
1
2 1
1
2 17
1
2 1
1
1
4 1
1 3
2 3
1
6
1 17
2
14
1 1
1 4
17
1 23
2
5
2
1 6
3
3 8
3 3 6 14 1 16 5 17 22 3 1 6 2 4 3 7 113
42
Table 3–4: Distribution of Acquiring and Target Firms by Industry and Year This table shows the distribution of block purchases by industry. Industry Crops Natural resource extraction Real estate development Heavy construction Food products Tobacco Apparel and textile mill products Furniture and fixtures Paper and allied products Printing and publishing Chemicals and allied products Rubber and plastics Leather and footwear Stone, clay, and glass products Primary products and metal Industrial machinery Electronic/electric equipment Transportation equipment Instruments and related products Local transit Water transportation Air transportation Miscellaneous transport Communications Utilities and waste management Wholesale Retail Services
SIC codes 1 10-14 15 16 20 21 22, 23 25 26 27 28 30 31 32 33, 34 35 36 37 38 41 44 45 47 48 49 50, 51 52-59 70-87
Number of target firms 1 2 5 3 3 2 1 3 6 3 2 4 5 6 2 4 1 2 1 1 13 12 3 4 24
Proportion 0.9% 1.8% 4.4% 2.7% 2.7% 1.8% 0.9% 2.7% 5.3% 2.7% 1.8% 3.5% 4.4% 5.3% 1.8% 3.5% 0.9% 1.8% 0.9% 0.9% 11.5% 10.6% 2.7% 3.5% 21.2%
Number of acquiring firms 1 7 1 4 1 1 1 3 5 9 2 2 5 1 5 1 6 1 1 1 4 10 12 5 5 19
Proportion 0.9% 6.2% 0.9% 3.5% 0.9% 0.9% 0.9% 2.7% 4.4% 8.0% 1.8% 1.8% 4.4% 0.9% 4.4% 0.9% 5.3% 0.9% 0.9% 0.9% 3.5% 8.8% 10.6% 4.4% 4.4% 16.8%
3.3.2 Factors that Explain the Value Creation from Corporate Block Equity Purchases To shed light on the determinants of wealth creation from corporate block purchases, we identify several cross-sectional factors that are closely related to the theoretical benefits of minority corporate block ownership. We also take into account factors that are not specifically limited to corporate block purchases, but which are likely to have an influence on the observed wealth effect.
43
Table 3–5: Variable Description This table contains the definitions and data sources for each variable collected. Variable name
Definition
Transaction characteristics Block size Proportion of the common stock of the target acquired Full takeover attempt Cross-border
Dummy variable taking the value of one if the acquirer attempted to take over the target within 1 year after the block purchase Dummy variable taking the value of one if target and acquirer have different home countries
Target industry characteristics High-tech industry Dummy variable taking the value of one if the target operates in a high-tech industry Regulated industry Dummy variable taking the value of one if the target operates in a regulated industry Product market relationship related characteristics Strategic alliance/ Dummy variable taking the value of one if the target joint venture operates in a regulated industry Industry relatedness Dummy variable taking the value of one if target and acquirer share the same primary two-digit SIC code Target financial characteristics Prior performance The buy-and-hold return of the target relative to the buyand-hold return of the specific country market index over the estimation period Market-to-book ratio The market-to-book ratio of the target at t-21 Market value The market value of equity of the target at t-21 Relative size The market value of equity of the target divided by the acquirer’s market value at t-21 Free cash flow Ratio of the free cash flow (Worldcope item ws.freecashflow) and total assets (Worldscope item 02999) of the target
Sources
SDC/ Thomson One Banker; Factiva Factiva
SDC/ Thomson One Banker
Hecker (1999) Campa and Hernando (2004)
Factiva SDC/ Thomson One Banker
Datastream
Datastream Datastream Datastream Worldscope
We classify these factors into four sets of characteristics: industry characteristics, product market relationship related characteristics, transaction characteristics, and target financial characteristics. In the following, we discuss these factors and identify related proxy variables for our empirical analysis. Table 3-5 gives an overview over the proxy variables and respective data sources used in this study.
44
3.3.2.1
Industry Characteristics
With regard to target industry characteristics, we hypothesize the technology intensity of the target industry to be an important factor with regard to the wealth creation potential of corporate block purchases. As already discussed, the mitigation of information asymmetries and misaligned incentives in business relationships can be regarded as a specific potential benefit of corporate minority block ownership. Given the inherent range for information asymmetries, contracting problems between corporations seem particularly likely in high-technology industries. Pisano (1989) argues that partial ownership may be especially effective for activities such as R&D that are costly to govern through contracts but are subject to incentive losses when internalized completely. This view is supported by Fee et al. (2006) who find that equity stakes are much more common in suppliercustomer relationships where the supplier is a R&D intensive firm. Accordingly, Allen and Phillips (2000) find that targets operating in industries with high R&D expenses show significant improvements in operating cash-flows and increases in investment expenditures following corporate equity block purchases, which they partly contribute to the mitigation of information problems between the corporations involved in the block purchase. Block purchases by an outside corporation may also help to mitigate information problems regarding the specific investment opportunities of target firms and validate the target’s investment opportunities to the capital markets. As the assessment of technologies can be highly complex and resource-consuming, evaluating the quality and potential of a high-tech firm is difficult for investors. Accordingly, empirical research indicates a relatively large information asymmetry associated with R&D. Barth, Kasznik and McNichols (2001) report that the number of analysts covering a firm is significantly larger for firms with intensive R&D. Aboody and Lev (2000) find that investors’ reaction to the public disclosure of insider trades is significantly stronger for R&D-intense companies, implying a larger information asymmetry in R&D-intense firms. In turn, the validation of the target company’s investment opportunities to the capital market via a block equity purchase by another corporation seems particularly relevant. Consequently,
45
we expect a positive impact on shareholder wealth for both the acquirer and the target if the target operates in a high-tech industry. In order to classify industries according to their R&D intensity, we use the classification scheme of Hecker (1999). This scheme identifies 31 three-digit SIC codes that comprise high-technology industries based on measures of industry employment in both R&D and technology-oriented occupations.20 We assign a dummy variable that is set to the value of one if the target company operates in a high-technology industry. Apart from the R&D intensity, also the regulatory environment of the specific industry may play a role with regard to value creation. Demsetz and Lehn (1985) argue that systematic industry regulation restricts the strategic options available to owners. In addition, they assume that regulation also provides some subsidized disciplining and monitoring of the management of regulated firms. In consequence, both the potential for value enhancing activities by a corporate blockholder (e.g. takeovers) and the incremental benefits from monitoring should decrease. Accordingly, Campa and Hernando (2004) find that intra-European M&A announcements in regulated industries generate less shareholder value than transactions in unregulated industries. Hence, one should also expect lower returns for corporate block purchases where the target operates in a regulated industry or an industry in which the involvement of state-owned enterprises is substantial. As there is no exact definition of what may be considered a regulated industry, we follow Campa and Hernando (2004) who identify seven industries in the EU that have traditionally been regulated or where the government has had a substantial ownership stake.21 We include a binary variable that takes on the value of one if the target operates in a regulated industry.
20
Hecker (1999) systematically identifies 31 industries in which the number of R&D workers and technology oriented occupation accounts for a proportion of employment that is at least twice the average of all industries included in the analysis. 21 Campa and Hernando (2004) consider minerals, primary metals, transportation, communication, electricity, gas, sanitary services and financial institutions as regulated industries in all EU member countries. More precisely, these industries correspond to the following two-digit SIC codes: 10, 13, 33, 40, 44-45, 48-49, 60-61, 80.
46
3.3.2.2
Product Market Relationship Related Characteristics
In contrast to institutional and individual investors, corporate acquires of ownership blocks can have a specific product market relationship with the target company, as business agreements, alliances or joint ventures between the acquirer and target can be reached. In this context, the purchase of an equity block may create value by mitigating potential information asymmetries and by aligning the incentives of the companies involved. Allen and Philipps (2000) present supporting evidence for this hypothesis by showing that the capital market reaction to the formation of corporate ownership blocks is significantly more positive if the parties involved in the transaction are already involved in an explicit partnership agreement such as a strategic alliance or a joint venture. Accordingly, Fee et al. (2006) find that target shareholder abnormal returns upon the announcement of minority equity purchases are significantly higher if the firms expressly state the strategic alliance as a principal consideration of the deal. Thus, one should expect higher abnormal returns for both the target and the acquirer in the presence of a strategic alliance or a joint venture. In order to control for this aspect, we identify those transactions that are accompanied by an existing or a new contractual business agreement such as a strategic alliance or a joint venture between the target and the block purchaser. We identify strategic alliance and joint venture agreements by searching the Factiva database for related news announcements over the entire one year period preceding the block purchase announcement for each of the 113 transactions.22 We assign a dummy variable that is set to the value of one if we can confirm the existence of a strategic alliance or joint venture. Even if there is no contractual relationship, the industry relatedness of the acquiring and target firm may influence the wealth effect from corporate minority block purchases. An acquirer that operates in the same product market as the target is likely to have a higher industry experience and more specific operational knowledge, thus creating a higher potential for effective monitoring and the realization of operational synergies. If the companies operate in an industry that is characterized by a concentrated and oligopo22
A further empirical distinction between newly formed and existing alliances and joint ventures does not change our empirical findings qualitatively.
47
listic market structure, the block purchase could also facilitate collusive behavior of the firms, which may increase the likelihood of increased monopoly rents. Hence, we hypothesize a positive effect of industry relatedness on the wealth effect for both target and acquirer. In order to account for industry relatedness of the two firms involved in the block purchase we follow Morck, Shleifer and Vishny (1990) and assign a dummy variable that is set to the value of one if both companies operate in the same two-digit SIC industry.
3.3.2.3
Transaction Characteristics
With regard to general transaction characteristics, we identify three factors that potentially have an effect on the capital market’s reaction upon the announcement of the block purchase. First, the size of the acquired equity block is likely to affect potential monitoring activities by the partial acquirer. Blockholders owning large stakes are more likely to monitor management actions since they face a liquidity problem (Maug (1998)). Moreover, the likelihood that the benefit of monitoring exceeds the cost increases with the size of the blockholding (Park et al. (2008)). Hence, we expect a positive relation between block size and abnormal returns. We include into our sample the proportion of target voting stock acquired (in %) as a measure for block size as reported in the SDC database. A second potential factor is the geographical focus of the transaction. A number of arguments underlines why cross-border transactions may create less value than national transactions. First and foremost, cross-border transactions are different to national transactions in that they usually combine two national cultures and two languages. Analyzing a sample of full takeovers, Denis, Denis and Yost (2002) argue that this should induce higher costs for post-merger integration. Although corporate minority block purchases do not necessitate physical and/or organizational integration processes, differences in national culture and language may aggravate acquirers’ monitoring activities and the ease of potential strategic collaborations. Moreover, the geographical distance between target and acquirer is usually larger in cross-border mergers than in national M&A transactions. Assuming that monitoring costs increase with physical distance, the incentive to effec-
48
tively monitor target management should be lower for remote acquirers. Hence, we expect cross-border block purchases to exhibit lower returns than national block purchases for both target and acquirer. We include a dummy variable that is set to one if the transaction is a cross-border block purchase. Third, the underlying transaction rationale may substantially affect the stock price reaction. If there is an obvious takeover intention, the observed wealth effect from these transactions is likely to partially reflect and anticipate change in majority control and the implied takeover premium (Mikkelson and Ruback (1985); Choi (1991)). We expect a significant positive relation between a takeover intention and excess returns to target shareholders. In turn, we would expect negative excess returns to the acquiring corporation. In order to account for a potential takeover anticipation effect in our empirical analysis, we proceed in two steps. As this effect is likely to empirically dominate other factors that relate to monitoring or the mitigation of information problems hypotheses, we already excluded from our sample those transactions that are made with a stated takeover intention or in the middle of a tender offer process (see section 3.3.1). However, the capital markets may anticipate a subsequent takeover even in the absence of takeover related information upon the block purchase. Barclay and Holderness (1991) and Sudarsanam (1996) empirically confirm this hypothesis by showing that the announcement effect upon a block purchase is much larger for target firms that subsequently experience a full change of control. This effect seems particularly relevant in the context of corporate block purchases, since block purchases by corporate bidders are more likely to result in a full acquisition than block purchases by other types of investors (Akhigbe, Martin and Whyte (2007)). Hence, we control for this potential anticipation by including a dummy variable taking the value of one if the acquirer makes a takeover bid for the target company in the one year period following the block purchase. We screen all related press announcements in the Factiva database after the announcement of the block purchase to identify takeover attempts by the block acquirer.
49
3.3.2.4
Target Financial Characteristics
Apart from industry characteristics, product market relationship characteristics and transaction characteristics, several financial characteristics of the target firm may reveal important insights into the potential effects from a corporate minority block purchase. First, the prior performance of the target firm is likely to affect the capital market’s evaluation of the potential benefits from a new corporate blockholder. On the one hand, poor stock price performance is likely to be correlated to management quality. According to this perception, several empirical studies find an inverse relationship between prior stock price performance and top management turnover (e.g. Warner, Watts and Wruck (1988), Denis and Kruse (2000)). Given the negative relationship between prior stock price performance and perceived management quality, the potential benefits from incremental monitoring by a new corporate blockholder should be inversely related to target firms’ prior performance. Consequently, we expect an inverse relation between abnormal returns and prior performance. We measure the historical stock price performance as the buy-and-hold return of the target firm’s stock relative to the specific index return over the estimation period. Furthermore, the target firms’ free cash flow may help to explain the observed wealth effect from corporate block equity purchases. The free cash flow hypothesis argues that firms with excess cash are likely to face significant agency costs if the excess cash is not distributed to shareholders. Barring such a distribution, managers have the incentive to invest the excess cash in empire building, perquisites and other projects with negative net present value (Jensen (1986)). Hence, the incremental monitoring benefit from a new corporate blockholder should be higher if the target has a relatively high cash flow indicating agency problems. Apart from the agency cost perspective, the role of free cash flow in motivating takeovers is frequently emphasized. Smith and Kim (1994) argue that acquisitions that relief free cash flow should generate higher gains to shareholders of bidders. Consequently targets with relatively high levels of free cash flow may be more attractive takeover candidates. Thus, we expect that the wealth effect is more positive if
50
the target has a high level of free cash flow. We include a variable that measures the ratio of free cash flow to total assets for the year end prior to the block purchase. Also the size of the target company may have an influence on the wealth effect of the transaction. Helwege et al. (2007) assume that the level of information asymmetries decreases with firm size. They argue that larger firms are more frequently monitored by analysts, institutional investors and regulators. We therefore expect lower incremental benefits from additional monitoring by a new corporate blockholder and thus a negative relation between market value and stock price reaction. We use the target’s market capitalization at t-21 as a proxy for firm size. Apart from the target’s absolute size, also the relative size of the target firm to the acquirer is likely to affect the wealth effects from a new block formation. However, the hypothesized direction of the effect is not clear a priori. On the one hand, if the acquirer is substantially larger that the target firm, the acquirer may be less likely to closely monitor the target firm’s management, since potential gains from effective monitoring are limited in absolute terms (Akhigbe et al. (2004)). Hence, the capital market should react more favorably when the size of the target relative to the acquirer is high, as effective monitoring becomes more likely. On the other hand, the probability of a subsequent change in majority control may diminish with increasing relative size of the target firm, since financing a takeover becomes increasingly difficult for the acquirer. In this case, the relative size of the target firm should be inversely related to the announcement returns. In our analysis, we use the ratio of the target’s market value and the acquirer’s market value at t-21 as a measure of relative size. Similar approaches are used by Servaes (1991), Mulherin and Boone (2000), and Shahrur (2005). Finally, stake purchases might be perceived as a signal that the target company is undervalued (Banerjee et al. (1997)). We therefore include the market-to-book ratio to proxy the target firms’ valuation levels. Table 3-6 presents descriptive statistics for the variables described above.
51
Table 3–6: Descriptive Statistics of Continuous and Binary Variables This table describes the data collected for the description of the transaction characteristics, target industry characteristics, product market relationship related characteristics and target financial characteristics. Proportions are reported for binary variables. Mean, median, standard deviation, minimum and maximum are reported for continuous variables. A description of the individual variables and the respective data sources is contained in Table 3.5. N Transaction characteristics Block size Full takeover attempt Cross-border Target industry characteristics High-tech industry Regulated industry
Proportion
Median
Stdev.
Minimum Maximum
113 12 47
10.6% 41.6%
16.9% -
15.0% -
9.9% -
5.0% -
45.4% -
26 33
23.0% 29.2%
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-1.7% 52.1% 1,997.7 7.2% 2.8
-1.2% 19.7% 291.4 8.0% 1.8
40.3% 80.4% 6,021.2 11.1% 5.4
-135.1% 0.1% 10.6 -45.1% 0.31
151.6% 500.5% 49,174.9 41.5% 53.7
Product market relationship related characteristics Strategic alliance/ 20 17.7% joint venture Industry relatedness 44 38.9% Target financial characteristics Prior performance Relative size Market value [in $m] Free cash flow Market-to-book ratio
Mean
113 113 113 113 102
-
The average acquired block size is quite substantial with a mean of 16.9% and a median of 15.0%. A full takeover attempt by the acquirer within 1 year after the block purchase is only observed in 10.6% of the transactions. As we already excluded transactions that took place in the middle of a tender offer process, subsequent to a takeover bid or with a stated takeover intention of the acquiring company, these takeovers should not have been strongly anticipated upon the announcement of the block purchase. With regard to the geographic scope of the transactions, cross-border block purchases constitute a large proportion of the total sample (41.6%). Concerning target industry characteristics, we find that 23.0% of the target companies operate in high-tech industries. The proportion of target companies that are from a regulated industry is slightly higher (29.2%). With regard to the financial characteristics of the target company, we find that target firms’ buy-and-hold return relative to the respective equity index was -1.7% during the estimation period on average. However, there seems to be a substantial disparity in the prior performance.
52
Analyzing the relative size of the target and acquiring firms, we find that relative market capitalization of target and acquiring firms at t-21 is on average 52.1%, with a median of 19.7%. This ratio is in line with those of majority acquisitions, as for example reported by Beitel, Schiereck and Wahrenburg (2004) for the European banking industry. The average market value of equity of the target companies is $1,997.7m at t-21 with a median of $291.4m. The average free cash flow ratio of the target firms is 7.2%. Finally, we document an average market-to-book ratio of 2.8, which documents the potentially attractive growth options of target companies as in the high-tech industry.
3.3.3 Method of Measuring Abnormal Returns To assess the value implication of the corporate block purchases, we use standard event study methodology. Expected returns are generated from the market model parameters, estimated with daily returns from 220 days before the block purchase announcement (t-220) to 21 days before the announcement (t-21). Adjusted prices (taking into account dividend payments and relevant changes to the capital structure) are used to calculate stock returns. As for the market return RM,t, we employ the Datastream price index in the companies’ home country. Abnormal returns (ARs) are calculated as the difference between actual returns and estimated returns from the market model in the event window. The capital market reaction is evaluated through different samples, in which mean values aggregate the abnormal returns per each observation day (Henderson (1990)). We use six different event windows ranging between 3 days [-1;1] and 41 days [-20;20]. Following the suggestion by Harrington and Shrider (2007), we use the test statistic of Boehmer et al. (1991) to test the significance of cumulated abnormal returns. The test statistic z is used to account for the likely difference in cross-sectional return variance between the estimation period from t-220 to t-21 and the event window. It is calculated as follows:
53
z
N
1 N
(1)
¦ SR
i ,t
i 1
N SR § 1 ¨¨ SRi ,t ¦ i ,t ¦ N ( N 1) i 1 © N i 1 N
· ¸¸ ¹
We calculate the standardization factor SRi,t following the approach of Mikkelson and Partch (1988): (2)
CARi , e1 e2
SRi ,t
¦ R e2
2
T t sˆi Ts s T
TS RM
2
M ,t
e1 T
¦ R
M ,t
RM
2
t 1
e1,e2 CARi ,e1 e2
N RM,t RM sˆi
SRi,t T Ts
: First day, last day in event window : Cumulated abnormal return of stock i in event window [e1, e2] : Number of stocks in sample : Return on market index on day t (event period) : Average return on market index in estimation period : Estimated standard deviation of abnormal returns for stock i : Standardized abnormal return on stock i at day t : Number of days in estimation period : Number of days in event window
The test statistic z follows a student t-distribution with T-2 degrees of freedom. The test results appear to be robust also in the absence of event-induced variance increases (Serra (2004)).
3.4
Results
3.4.1 Univariate Analysis Panel A of Table 3-7 presents the abnormal announcement period returns for the full sample of corporate minority block purchases. We find that all significant wealth gains in these transactions accrue to the shareholders of target firms. Abnormal returns to block acquirers are low (between 0.41% and 1.36%) and statistically not significant, perhaps
54
because the acquiring firms are, on average, substantially larger than the target firms. When we combine the abnormal returns by weighting each firm by its market capitalization at t-21, we find that the combined abnormal returns are between 1.05% and 1.95% and statistically significant for five out of six event windows. Given these findings, we focus predominantly on the impact of corporate equity purchases on target firms in subsequent discussions of the wealth effects. However, we note that the abnormal returns to target firms and the acquirers in the subsamples reported in Panel B of Table 3-7 are of the same sign for all but two classifications. We also test for the possibility of wealth transfer between the target and the acquirer by estimating the correlation between the cumulative average abnormal returns. If the acquirer was benefiting at the expense of the target or vice versa, we would expect a statistically significant negative correlation. However, the resulting correlation coefficients are statistically not different from zero.23 Thus, we find no evidence for wealth transfers between targets and acquirers.
23
This finding applies to all of the six different event windows used.
55
Table 3–7: Excess Returns around Block Equity Purchases by Corporations This table shows the average cumulative abnormal returns for the total sample of 113 minority block purchases in panel A and for subsamples in panel B. Only the CAARs for the [-5;1] event window are reported in panel B. Statistical significance at the 1%, 5%, and 10% level is denoted with ***, **, *. The statistical significance is tested using the test-statistic of Boehmer, Musumeci and Poulsen (1991) (zstatistic). Panel A: Total sample Event time period [-10;1]
Target CAAR z-statistic 7.44% 6.36***
Acquirer CAAR z-statistic 1.20% 0.96***
Combined Entity CAAR z-statistic 1.93% 2.48***
[-5;1]
6.61%
6.45***
0.78%
0.63***
1.73%
2.61***
[-1;1]
5.44%
5.80***
0.41%
0.07***
1.21%
1.91***
[-5;5]
6.27%
5.70***
1.30%
1.44***
1.95%
3.10***
[-10;10]
5.90%
4.24***
0.97%
0.99***
1.05%
1.77***
[-20;20]
7.74%
4.25***
1.36%
1.03***
1.29%
1.42***
Panel B: Subsamples Subsample Transaction characteristics Takeover attempt (N = 12) No takeover attempt (N = 101) Cross-border deal (N = 47) National deal (N = 66)
CAAR
Target z-statistic
Acquirer CAAR z-statistic
Combined Entity CAAR z-statistic
9.26%
1.21***
-2.62%
1.34***
-0.11%
0.73***
6.30%
6.68***
1.18%
1.38***
1.95%
3.18***
7.34%
4.80***
1.10%
0.85***
2.05%
2.18***
6.09%
4.45***
0.55%
0.10***
1.50%
1.61***
2.71***
0.94%
0.74***
1.04%
0.84***
5.86***
0.73%
0.45***
1.93%
2.47***
3.45***
0.88%
0.50***
1.58%
1.354***
5.44***
0.74%
0.38***
1.79%
2.25***
Target industry characteristics High-tech industry 6.13% (N = 26) Low- or mid-tech industry 6.76% (N = 87) Regulated industry 5.31% (N = 33) No regulated industry 7.15% (N = 80) (continued)
56
Table 3-7: Continued Target Subsample
CAAR
Acquirer
z-statistic
Product market relationship related characteristics Strategic alliance or joint 4.41% 2.53*** venture (N = 20) No strategic alliance or joint 7.09% 5.91*** venture (N = 94) Related industries 6.54% 3.28*** (N = 44) Unrelated industries 6.66% 5.67*** (N = 69) Both companies operate in 2.40% 0.48*** gas, electricity or water sector (N = 9) Companies operate in sec6.98% 6.55*** tors other than gas, electricity and water (N = 104)
Combined Entity
CAAR
z-statistic
CAAR
z-statistic
2.90%
1.76***
3.27%
2.13***
0.32%
0.02***
1.40%
2.03***
2.03%
1.15***
2.94%
2.19***
-0.02%
0.15***
0.96%
1.64***
-0.19 %
0.58***
0.39%
0.59***
0.86%
0.46***
1.85%
2.53***
Figure 3-1 illustrates a dynamic view on the development of the cumulative abnormal returns over the [-10;10] event window. The results show that there is a run-up in the targets’ average abnormal returns before the event. There are two possible explanations for this observation. Figure 3–1: Cumulative Average Abnormal Returns This figure illustrates the development of cumulative average abnormal returns around the announcement date at t0 over the period from t-10 to t+10 for target firms, acquiring firms, and the combined entity. Combined entity returns are calculated as the sum of the value weighted returns of the target and the acquirer, based on the market capitalization at t-21.
Cumulative average abnormal return
8%
Acquirer Target
6
Combined Entity
4
2
0
-10 -9
-8
-7
-6
-5
-4
-3
-2
-1
0
1
2
Days relative to announcement
3
4
5
6
7
8
9
10
57
First, a large amount of the block purchases may be carried out through open market purchases, and it can take months to accumulate a stake for which an official notification is required. As a consequence, the market may already react to increased buy-side pressure before a public announcement of a new block formation. Second, there is a potential lag between the actual block purchase and the public announcement. Hence, it cannot be ruled out that information on the transaction is already known to individual market participants prior to the public announcement date. Because of the possibility that stock market participants knew about the pending stake before it was announced, we focus on the [-5;1] event window in the following discussion.24 In the full sample, we find that target firms experience a highly significant average abnormal return (5.9% to 7.4%) upon the announcement of a block equity purchase by an outside corporation. This results is highly consistent with Allen and Phillips (2000) who report a cumulative abnormal return of 6.1% over the [-10;10] event window. Compared to other studies that do only include other types of acquirers such as banks or trusts, abnormal returns in our study appear higher. For example, Mikkelson and Ruback (1985) find for interfirm investments an excess return of 3.24% over the [-1;0] event window,25 while Wruck (1989) reports an excess return of 4.32% over the [-5;1] event window. A possible explanation for this observation is that corporate block purchases bear, on average, a higher potential for value creation, given the specific benefits discussed in section 3.2.1. Analyzing subsets of the total sample with regard to transaction characteristics presented in Panel B of Table 3-7, we find that targets experiencing a takeover bid within the year after the announcement of the block purchase show a tendency to higher abnormal returns (9.26%) than targets that are not subject to a takeover bid in the subsequent year (6.30%). In turn, acquiring companies that make a takeover bid for the target within the subsequent year experience lower abnormal return than acquirers who remain a mi24
Using the average cumulative abnormal return over the [-10;1] event window as the dependent variable leads to the same qualitative results. 25 This result is based on a subsample of 106 block purchases by other firms who stated that their intention was a financial investment only.
58
nority shareholder and do not try to gain majority control over the target (-2.62% vs. 1.18%). Although these differences in excess returns are not significant at conventional levels,26 we interpret this finding as indicative for the ability of the capital market to anticipate a full takeover and thus price in the anticipated takeover premium for the target firm. Concerning the geographical scope of the transaction, we do not find a significant difference between the abnormal returns of national and cross-border transactions. In order to estimate the effect of target industry characteristics on the excess returns, we also calculate cumulative abnormal returns for corporate block purchases by whether target firms operate in high-tech or regulated industries. In line with our initial hypothesis, average abnormal returns are higher for targets that operate in industries that are not regulated or dominated by state-owned companies (7.15% vs. 5.31%). In contrast, we find that the excess returns to target firms that operate in low-tech industries are slightly lower than those to target firms that operate in high-tech industries (6.13% vs. 6.76%), which does not support our initial hypothesis concerning the high potential of corporate block ownership to resolve problems stemming from information asymmetries in hightech industries. However, we have to note that these differences in the cumulative abnormal returns are statistically not significant. With regard to the product market relation between target and acquirer, we only find a minimal difference in average abnormal returns to target firms when we split the sample by whether the companies involved operate in related or unrelated industries (6.66% vs. 6.54%). Although our results might indicate that acquiring companies benefit more from non-diversifying transactions than from transactions in unrelated industries (2.03% vs. -0.02%), we cannot document a significant mean difference in abnormal returns between the two subsamples.27
26
We conduct a standard mean-difference test for each of the pairs of subsamples discussed. We consider 1%, 5% and 10% as conventional levels of statistical significance. 27 In order to control specifically for the potential collusive effects of minority block purchases in highly concentrated industries, we also analyze an additional subsample of intra-industry block purchases in companies operating in the gas, electricity or water industry, since these industries are still likely to have an oligopolistic structure that is prone to collusive behavior. We cannot, however, find a significant difference in abnormal returns.
59
Remarkably, the average abnormal return to target firms has a tendency to be lower in the presence of new or existing alliance or joint venture agreements (4.41% vs. 7.09%). Although we do not find statistical mean difference at conventional levels, this result appears contrary to Allen and Phillips (2000) and Park et al. (2008) who document significantly higher abnormal returns to target shareholders upon block purchases that are accompanied by strategic alliances or joint ventures. However, we note that the empirical evidence with regard to the interaction effect of block ownership and strategic corporate agreements appears inconsistent. Häussler (2006) reports significantly lower cumulative abnormal returns upon announcements of strategic alliance if the acquirer at the same time invests in an equity stake of the target. Summing up, the univariate analysis of excess returns upon the announcement of corporate block equity purchases reveals a significant value generation for target shareholders and gives instructive preliminary indications concerning the factors that may determine the extent of the effect. To further extend our analysis we conduct multivariate analyses in the following section.
3.4.2 Cross-Sectional Regression Analysis To gain additional insights in the determinants of the cumulative average abnormal returns to targets and acquirers following corporate block purchases, we conduct a crosssectional analysis. We estimate regressions using the cumulative abnormal return over the [-5;1] event window as dependent variable.28 Table 3-8 and Table 3-9 report the regression specifications for the cumulative abnormal returns to targets and acquirers, respectively. All test statistics are computed using White’s heteroskedasticity-consistent covariance matrix (White (1980)).
28
Using the average cumulative abnormal return over the [-10;1] event window as the dependent variable leads to the same qualitative results.
60
Table 3–8: Determinants of Cumulative Average Abnormal Target Returns This table shows estimation results for OLS-regression models with the cumulative abnormal return for the [-5;1] interval as the dependent variable. The variables and data sources are laid out in Table 3-5. Tvalues are in brackets. Statistical significance at the 1%, 5%, and the 10% level is denoted with ***, **, *. All test statistics are computed using a heteroskedasticity-consistent covariance matrix (White (1980)). No evidence for multicollinearity was found. Model 1 Intercept
Model 2
0.006)*** (0.260)***
0.076)*** (5.519)***
Model 3 0.071)*** (5.240)***
Model 4 0.164)*** (3.957)***
Model 5
Model 6
0.114)*** (2.376)***
0.119)*** (2.305)***
Transaction characteristics Block size
0.003)*** (2.361)***
0.002)*** (1.998)***
0.003)*** (2.105)***
Full takeover attempt
0.003)*** (0.185)***
-0.003)*** (-0.079)***
-0.020)*** (-0.557)***
Cross-border
0.017)*** (0.930)***
0.017)*** (0.920)***
0.019)*** (0.944)***
Target industry characteristics High-tech industry
-0.015)*** (-0.497)***
-0.023)*** (-0.872)***
-0.025)*** (-0.866)***
Regulated industry
-0.023)*** (-1.160)***
-0.026)*** (-1.239)***
-0.029)*** (-1.401)***
Product market relationship related characteristics Strategic alliance/ joint venture
-0.027)*** (-1.019)***
-0.015)*** (-0.618)***
-0.007)*** (-0.272)***
Industry relatedness
-0.002)*** (-0.093)***
0.008)*** (0.389)***
0.006)*** (0.279)***
Target financial characteristics Market-to-book ratio
-0.001)*** (-0.923)***
Prior performance
-0.054)*** (-1.949)***
-0.050)*** (-2.052)***
-0.046)*** (-1.752)***
Log relative size
-0.168)*** (-2.738)***
-0.148)*** (-2.649)***
-0.137)*** (-2.285)***
Log market value
0.006)*** (0.489)***
0.007)*** (0.594)***
0.002)*** (0.159)***
Free cash flow ratio
0.252)*** (2.841)***
0.234)*** (2.338)***
0.248)*** (2.331)***
N
113
113
113
113
113
102
Adj. R2
6.4%
-0.1%
0.0%
13.7%
15.0%
13.3%
We do not find evidence on the presence of multicollinearity in any of the models based on the analysis of variance inflation factors. The variables included in our models are structured following our initial hypotheses on the factors that potentially explain the observed wealth effect.
61
We design six different regression models. In models 1 to 4, we test each set of characteristics separately. Model 5 includes all variables but the market-to-book ratio, since information is not available for all of the 113 transactions in our sample. Model 6 also includes the market-to-book ratio.29 As target corporations experience significantly higher and more significant abnormal returns than acquiring companies, we will first analyze the multivariate regression results of cumulative average abnormal returns to the target firms. With regard to the determinants of the abnormal returns, we do not find that target industry characteristics and product market relationship related characteristics help to explain the abnormal returns to target shareholders. The coefficients of the included variables are insignificant in all regression models. With regard to the R&D intensity of the target company, this confirms the result of our univariate analysis, which is consistent with evidence of Allen and Phillips (2000). In contrast, we cannot confirm their univariate finding that transactions accompanied by a strategic alliance or joint venture have significantly higher excess returns than transactions where the corporations do not enter or already have formal agreements. As the strategic alliance/ joint venture variable is insignificant not only in model 2 where we only analyze the subset of variables relating to product market relationship of the companies, but also in models 5 and 6, we assume that our finding is not affected by cross-sectional differences in firms size or other transaction and/or target financial characteristics. Remarkably, the industry relatedness of the companies involved in the transaction does not have a significant effect. Given our initial hypothesis that industry relatedness does not only enhance monitoring due to the acquirers’ industry and operational expertise, but that it is also a precondition for collusive behavior, we would have expected a strong positive relationship. Eventually, this finding indicates that the potential positive effect of relatedness does not fully accrue to the target company, but to the block acquirer. 29
Although the market-to-book ratio is available for all targets, we do not include 9 observations with negative values in our analysis.
62
Considering the transaction characteristics, we find a significant positive influence of the acquired block size on abnormal return. This finding is consistent with our initial hypothesis that the size of the block is positively related with the likelihood of effective monitoring by the new blockholder. The geographical scope and a subsequent takeover bid do not influence the regression results significantly. Compared to the other sets of variables, the financial characteristics of the target have the highest explanatory power. The signs of all three statistically significant variables are consistent with our initial hypotheses. Targets that performed poorly over the 200 days of the estimation period relative to the respective benchmark index experienced significantly higher abnormal returns. This result is consistent with Akhigbe et al. (2004) who report a significant negative relation between the announcement returns of block purchases and the prior stock price performance. As stock price performance is frequently related to management quality, a possible explanation for this finding is that the capital market may expect a performance enhancement via more effective monitoring by a new corporate blockholder. We also find an inverse relation between the level of free cash flows and abnormal returns. As agency conflicts between shareholders and managers are hypothesized to increase with free cash flows, the capital market may attribute a higher potential for an effective mitigation of agency problems by the new outside investor if the target company has relatively high free cash flows. The size of the target relative to the acquirer has a negative relation to the abnormal returns. Although the acquiring company may have a lower incentive to effectively monitor management if the target is substantially smaller (Akhigbe et al. (2004)), the probability of a subsequent change in majority control is likely to increase, since financing a takeover becomes increasingly feasible for the acquirer. The latter reasoning seems to be reflected in the stock price reaction. Overall, these results suggests that the positive stock price reaction for target companies is not significantly driven by factors that specifically relate to the theoretical benefits of corporate equity ownership.
63
The multivariate analysis of the average abnormal returns to the acquiring firms reveals the same picture. With regard to target industry characteristics and product market relationship related characteristics, we only find industry relatedness to have a significant effect on the observed stock price effect. As the acquiring companies in our sample are significantly larger than the target companies (median of relative size = 19.7%), we do not assume that this effect is solely due to better monitoring. Rather, as industrial diversification is associated with a significant value discount (Denis et al. (2002)), one should expect a more positive wealth effect from transactions that take place in a related industry. Another potential explanation is that an acquiring company operating in a related industry is able to capture the positive effect from potential collusive behavior. However, as the potential for collusive behavior is probably limited to a small number of transactions in our sample, we attribute our finding to the capital markets’ negative assessment of diversifying transactions. With regard to the financial characteristics, we find three factors that have a significant influence on abnormal returns. Apart from the size of the target, both the prior performance and the free cash flow of the target have a significant impact on the cumulative abnormal returns of the acquirer. The signs of the coefficients are the same as in the regression models of the target cumulative abnormal returns. This suggests that value created from effective monitoring and the mitigation of free cash flow related agency problems accrues to both target and acquirer.
64
Table 3–9: Determinants of Cumulative Average Abnormal Acquirer Returns This table shows estimation results for OLS-regression models with the cumulative abnormal return for the [-5;1] interval as the dependent variable The variables and sources are laid out in Table 3-5. T-values are in brackets. Statistical significance at the 1%, 5%, and the 10% level is denoted with ***, **, *. All test statistics are computed using a heteroskedasticity-consistent covariance matrix (White (1980)). No evidence for multicollinearity was found. Model 1 Intercept
Model 2
0.008)*** (0.553)***
0.006)*** (0.606)***
Model 3 -0.005)*** (-0.647)***
Model 4 0.037)*** (1.429)***
Model 5
Model 6
0.031)*** (0.923)***
0.042)*** (1.081)***
0.000)*** (0.144)***
-0.000)*** (-0.257)***
-0.000)*** (-0.655)***
-0.039)*** (-1.495)***
-0.039)*** (-1.528)***
-0.030)*** (-1.002)***
0.006)*** (0.432)***
0.005)*** (0.411)***
0.005)*** (0.349)***
Transaction characteristics Block size Full takeover attempt Cross-border Target industry characteristics High-tech industry
0.003)*** (0.816)***
-0.008)*** (-0.554)***
-0.003)*** (-0.196)***
Regulated industry
0.002)*** (-0.138)***
-0.002)*** (-0.124)***
-0.005)*** (-0.321)***
Product market relationship related characteristics Strategic Alliance/ Joint Venture
0.027)*** (1.593)***
0.024)*** (1.401)***
0.028)*** (1.412)***
Industry relatedness
0.021)*** (1.565)***
0.027)*** (2.131)***
0.029)*** (2.217)***
Target financial characteristics Market-to-book ratio
-0.000)*** (-0.518)***
Prior performance
-0.035)*** (-2.266)***
-0.033)*** (-2.132)***
-0.035)*** (-2.150)***
Log Relative size
0.016)*** (0.488)***
0.016)*** (0.454)***
0.008)*** (0.230)***
Log Market value
-0.017)*** (-2.258)***
-0.019)*** (-2.518)***
-0.020)*** (-2.284)***
0.013)*** (0.323)***
0.077)*** (1.493)***
0.092)*** (1.680)***
Free cash flow ratio N
113
113
113
113
113
102
Adj. R2
0.0%
-1.8%
2.5%
4.5%
6.1%
5.4%
Summing up, our results suggest that the stock price effect upon the announcement of minority corporate block purchases does not reflect the specific potential benefits of corporate equity ownership such as the mitigation of information problems and the alignment of incentives. In turn, we find that factors that signal monitoring or agency
65
problems such as the prior performance of the target or the level of free cash flows have a significant influence on the abnormal returns to both targets and acquirers.
3.5
Conclusions
In this study, we investigate share price reactions to corporate minority block purchases for a sample of 113 European block purchases between 1993 and 2006. We argue that corporate block ownership has several unique characteristics relative to institutional or individual blockholders. For this reason, we identify four sets of characteristics that relate to the main hypotheses with regard to the potential sources of wealth creation. We then ask whether these have a differential significant impact on abnormal announcement returns to target firms and block purchasers. In the first part of our empirical analysis, we document a significant wealth creation for target companies. The finding of significant positive announcement returns to target firms is consistent with Mikkelson and Ruback (1985), Wruck (1989), and Allen and Phillips (2000). Although we cannot report significant abnormal returns to corporate block purchasers, we find the abnormal returns to the combined entity to be significantly positive for five out of six event windows. However, we do not find significant differences in wealth creation with regard to either industry characteristics or product market relationship related characteristics. Consistent with our univariate results, we do not find that industry characteristics and product market relationship related characteristics play an important role in explaining abnormal returns to targets and acquirers when we control for other cross-sectional factors. In addition, our analysis reveals that financial characteristics of the target firms such as the prior stock price performance or the level of free cash flows have the highest explanatory power. We show that the stock price reaction for both target and acquirer is significantly more positive if the target shows indications of poor management performance and/or agency problems.
66
In conclusion, our results may support two different arguments considering our initial hypotheses with regard to the specific theoretical benefits of corporate block ownership. On the one hand, under the assumption of efficient capital markets, our results suggest that the specific benefits of corporate block ownership such as the mitigation of information asymmetries in business relationships are limited in size and hence not significantly reflected in the stock price reaction. On the other hand, one could assume that the capital market does not instantly and correctly factor in these potential benefits, as they are rather subtle and difficult to measure. Validations of this interpretation in further detail should be a fruitful topic for future research.
67
4
Minority Equity Ownership and Value Creation: The Role of Corporate Relatedness
4.1
Introduction
In contrast to takeovers and acquisitions, minority block purchases do not give the acquirer full control over the target firm. Hence, one should suppose that their potential for wealth generation is substantially lower. As new minority blockholders cannot exert full control over the target firm, the realization of wealth creating strategies such as the implementation of operating synergies or increased market power should be inherently limited compared to acquisitions of majority control stakes. Despite these considerations, non-financial corporations have been active acquirers of long-term equity positions in U.S. firms (Allen and Phillips (2000), Cronqvist and Fahlenbrach (2007)). Empirical research indicates that acquisitions of minority equity blocks can lead to significant increases in shareholder value. Part of this wealth creation seems to stem from an anticipated takeover premium in the context of “toehold” acquisitions (e.g. Mikkelson and Ruback (1985)). More recent research also highlights the potential mitigation of agency conflicts as an important value driver in minority equity purchases. For example, Allen and Phillips (2000) find that corporate block ownership leads to a larger increase in shareholder wealth, investment, and operating profitability if it is combined with joint ventures, alliances, and other product market relationships, due to the higher potential for the mitigation of information asymmetries and the alignment of incentives in this context. Above all, these observations indicate that the potential for wealth seems to depend on firm- and industry-specific characteristics. With regard to firm-specific characteristics, related empirical research on takeovers (i.e. acquisitions of majority control) suggests that corporate relatedness between the acquirer and the target may play an important role in understanding why some transactions create value and others do not. Related takeovers generate more operating synergies than diversifying mergers (e.g. Haely et al. (1992), Maquieira et al. (1998), Maksimovic and Phillips (2001)). Also the type of relatedness seems to affect the potential for value crea-
68
tion. For example, vertically related mergers seem to generate significantly greater positive wealth effects than vertically unrelated mergers (Fan and Goyal (2006)). Although the relevance of corporate relatedness has yet not been empirically analyzed in the context of minority block ownership, 30 there are numerous theoretical considerations that lead us to the hypothesis that the potential benefits of minority block ownership strongly depend on the type of corporate relatedness between the block purchaser and the target. On the one hand, minority equity ownership between vertically related firms may foster vertical integration, which in turn can reduce transaction costs that arise due to contractual inefficiencies in customer-supplier relationships (e.g. Klein et al. (1978), Williamson (1979), Grossman and Hart (1986), Hart and Moore (1990)). In addition, the market power of the vertically integrated firm may increase (Salinger (1988), Hart and Tirole (1990), Ordover et al. (1990)). On the other hand, corporate equity ownership in a horizontally related firm may give incentives to reduce competition and foster collusive behavior (Reynolds and Snapp (1986), Malueg (1992)). Considering these arguments, we want to provide empirical evidence on the relationship between corporate relatedness and the wealth effects of minority block purchases. Our methodological approach accounts for the complexity of corporate relatedness. Although recent studies already try to measure relatedness based on primary SIC codes (Allen and Phillips (2000), Akhigbe et al. (2004)), their approach only reveals the horizontal relatedness of firms operating in the same primary industry. This simple and widely used measure, however, does not account for more complex types of relatedness such as vertical relatedness and complementarity. E.g., in 2001, the film studio MetroGoldwyn-Mayer Inc. (SIC code 7812; motion pictures and TV production) acquired a stake in Rainbow Media Corp. (SIC code 4841; cable and other pay television services) in order to gain access to own cable channels to which it can feed its movie produc30
We distinguish between three types of relatedness: (1) horizontal relatedness (2) vertical relatedness and (3) complementarity. We classify two businesses as horizontally related if they operate in the same industry. We define those businesses as vertically related that can use the other’s services or products as input for its own production or supply output as the other’s input. We classify two businesses as complementary if they have a high overlap in the markets they source their input from or the markets they sell their products to.
69
tions.31 Although both businesses are apparently vertically related, traditional measures of corporate relatedness used in recent studies of corporate block equity purchases would classify this transaction as unrelated, as they do not share the same primary SIC code. Even if two firms do not operate in horizontally or vertically related businesses, they may source their inputs from the same markets and sell their outputs to the same customer segments. In 2000, Johnson Controls Inc. (SIC code 3714; motor vehicle parts), which produces automobile batteries and interior components, acquired a stake in Donelly Corp. (SIC code 3231; glass products made of purchases glass), a producer of automotive glass products such as mirrors.32 Although these businesses are obviously not vertically integrated, they have a high overlap in the industries they sell these products to and thus increased potential to cooperate, e.g. through joint distribution efforts. These practical examples illustrate that the theoretical considerations with regard to the specific type of relatedness should be more thoroughly reflected in empirical research. Consequently, we follow the methodology of Fan and Lang (2000) and use commodity flow data from U.S. input-output (IO) tables provided by the U.S. Bureau of Economic Analysis to construct quantitative measures of relatedness. Based on this methodology, we examine a sample of 141 minority block purchases during the 1984 to 1997 period, including only transactions where both the acquirer and the target are non-financial corporations based in the U.S. To the best of our knowledge, this is the first study that explicitly analyses the effects of relatedness in the context of corporate block ownership. Our approach to this study is as follows. We first present a brief overview of the related theoretical and empirical research. We then provide descriptive statistics on our sample. Next, we explain the methodology of measuring relatedness for the pairs of firms in our sample. We then examine the cumulative abnormal return to target firms, acquiring firms and the hypothetical combined entity around the announcement of block purchases. We conduct both univariate and multivariate analyses in order to shed light on
31 32
As reported in the Wall Street Journal on February 2, 2001. As reported in the Wall Street Journal on August 11, 2000.
70
the relationship between corporate relatedness and the observed wealth effects. The most important results of this study can be summarized as follows: More than half of the minority block purchases are either between corporations in the same industry (34.8%) or between firms with significant vertical relatedness and/or complementarity (24.8%). We find abnormal returns to target firms and abnormal combined value-weighted entity returns to be positive and significant. This indicates that the capital market perceives the formation of a corporate equity position as value creating. Based on our measures of relatedness, we find a positive but statistically weak relationship between the presence of relatedness and the observed wealth effects. This may be due to either the limited economic magnitude of the potential benefits that stem from equity ownership between firms in related industries or to a lagged or incomplete valuation of potential economic benefits by capital markets. These results also hold when we expand our analysis to secondary segment information in order to account for potential effects of relatedness across primary and secondary business divisions. The remainder of this paper is structured as follows. In section 4.2 we discuss theoretical causes and consequences of corporate block purchases and review the related empirical literature. Section 4.3 describes the sample selection process and our research methodology for measuring abnormal returns and for constructing the relatedness coefficients. Section 4.4 contains our empirical results and section 4.5 concludes.
4.2
Hypotheses and Related Literature
In this section, we discuss possible reasons for equity ownership between firms and briefly review the related theoretical research. We focus on the motives that are specifically related to the scope (i.e. vertical, horizontal, complementary) of the ownership agreement and form initial hypotheses with regard to their potential wealth effects.33 If a company acquires an equity stake in another company that either operates in a customer (downstream) or supplier (upstream) industry, a potential motive may be to 33
A more general literature review on the potential benefits of corporate block ownership is provided by Allen and Phillips (2000).
71
increase vertical integration. Theoretical research suggests that vertical integration can increase efficiency in corporate relationships through the reduction of transaction costs by mitigating contractual frictions and through the alignment of incentives. There are two leading theoretical approaches that explain the potential motives and benefits of vertical integration between two corporations. The transaction cost theory (Coase (1937), Williamson (1979), Klein et al. (1978)) views vertical integration as a way of circumventing contractual inefficiencies (“holdup”), and predicts that vertical integration should be most beneficial when there is greater asset specificity and uncertainty in market transactions. The property rights theory (Grossman and Hart (1986), Hart and Moore (1990)) focuses on the role of ownership of assets as a ways of allocating residual rights of control. Hence, it emphasizes both the costs and the benefits of vertical integration in terms of ex ante investment incentives. Accordingly, Aghion and Tirole (1994) model several cases in which the optimal solution, given relationship-specific investments by both parties, may be partial ownership by a downstream firm of an upstream firm. Consistently, Fee et al. (2006) find that equity stakes are much more common in supplier-customer relationships where the supplier is a R&D intensive firm facing a high degree of information asymmetry. Apart from the efficiency-related motives discussed above, vertical integration may also be used to increase market power and foster collusive behavior. On the one hand, a vertical integration with a supplier (customer) could enhance the market power of the integrated entity since it can deny access of the input (output) to its non-integrated rivals (Salinger (1988), Hart and Tirole (1990), Ordover et al. (1990)). On the other hand, vertical integration with a supplier (customer) may enable better coordination between the integrated firm and its non-integrated acquirer rivals since the rivals need access to the input (output) being controlled by the integrated entity (Chen (2001), Nocke and White (2007)). Given the above discussions, we hypothesize that block ownership in the presence of vertical relatedness is likely to have a positive wealth effect due to potential increases in productive efficiency and market power.
72
With regard to horizontal partial equity ownership between corporations, block acquirers that operate in a horizontally related industry may monitor the target more effectively and have more opportunities to realize operational and financial synergies (Spencer et al. (1998), Akhigbe et al. (2004)). Furthermore, Reynolds and Snapp (1986) show that partial ownership agreements may reduce competition if the firms involved are actual or potential competitors. Accordingly, Malueg (1992) proves that increases in cross-ownership may lead to less competition and an increased likelihood of collusion. Consequently, we hypothesize a positive wealth effect of equity block purchases between horizontally related firms. Even if the firms operate in different businesses that are not vertically related, a high overlap in the markets they source their input from and/or the markets they sell their products to may inherent potential benefits for corporate block ownership. For example, corporate equity ownership may align the incentive of two complementary firms in joint sourcing or distribution initiatives. Moreover, equity participation may foster the exchange of proprietary information with regard to relevant input or output markets. Hence, we expect a positive wealth effect of corporate equity ownership if complementarity between two firms is high. In sum, prior research suggests that the presence and type of relatedness should be an important factor in explaining the wealth effects from corporate minority equity purchases. In the following, we provide empirical evidence on the relevance of industry relatedness in the context of corporate block ownership.
4.3
Data and Methodology
4.3.1 Corporate Block Purchases We analyze a sample of minority block purchases where both the target and the acquirer are U.S. non-financial companies. Relevant transactions were identified using the SDC
73
database. The initial sample comprises all 731 completed minority block purchases with public non-financial34 target and acquirer in the 1984 to 1997 period.35 We then use four criteria to further specify the sample: (1)
The acquirer owns 5% to 49.9% of the target after the transaction, and the acquirer does not have substantial (i.e. more than 5%) prior ownership
(2)
Acquirer and target are both public companies and daily stock prices for the period t-220 to t+20 are available in the Datastream database and neither the acquirer nor the target have more than 75% zero-return days during the estimation period
(3)
Details of the block purchase reported in the SDC database are confirmed by searching the Factiva database
(4)
Announcement of the block purchase does not take place in the middle of a tender offer process, subsequent to an announced takeover bid or with an expressly stated takeover intention
Following Choi (1991), we define outside minority blocks as investors who own more than 5% but less than 50% of the target firm’s voting stock after the purchase (criterion (1)). 36 Criterion (2) ensures that we only include transactions where sufficient return data is available for the estimation of abnormal returns for both target and acquiring firms via the market model. Applying criteria (1) and (2), we derive 204 transactions. We then try to further improve the data quality of our sample by confirming each of the remaining transactions reported in the SDC database. For this purpose, we screen the Factiva database for related information. We find information on 198 transactions (crite34
We exclude any transaction with either a financial acquirer or target (primary SIC code between 6000 and 6999) from our initial sample. 35 The reason for choosing the 1984 to 1997 period is twofold: First, historical segment information is only available from 1984 onwards on the Compustat database. Second, consistent SIC code based Use tables (see section 4.3.2) are only available for the years 1982, 1987 and 1992. We refrain to use the 1992 for transactions after 1997, given the potential changes in inter-industry commodity flows. Kale and Shahrur (2007) also apply the data from the Use tables for up to six years. 36 We also conduct our analyses for block purchases that lead to 5-30% ownership and obtain similar results.
74
rion (3)). Next, we further exclude 23 transactions that take place in the middle of a tender offer process, subsequent to a full takeover bid or with a stated takeover intention (criterion (4)), since the wealth effect from these transactions is likely to partially reflect an anticipated change in majority control and not the wealth effects stemming from relatedness-specific benefits. We use information obtained from the Factiva database to identify the transactions subject to this criterion. In order to maximize the reliability of our sample, we also use information obtained from the Factiva database to control for potential misspecifications in the SDC database. In total, we have to manually exclude 34 transactions due to misspecifications in the SDC database.37 Our final sample contains 141 transactions that fulfill the above criteria. Table 4-1 gives a detailed overview over the actual sample selection process. Table 4-2 includes the distribution of transactions by year. Table 4-3 gives an overview over the industry classification of both acquirers and targets, based on historical primary SIC codes. We find chemicals and allied products, electronic equipment, instruments and related products, and services to be the industries that constitute the largest number of both target and acquiring firms.
37
The 34 manual exclusions were made due to the following reasons: Announcement of concurrent material news (e.g. major investments or divestiture decisions, earnings surprises) (13), existence of substantial ownership stake (>4.9%) not reported in SDC database (9), block purchases leading to defacto majority control (5), transactions within one corporate entity (3), transactions with double entry in SDC database (4).
75
Table 4–1: Sample Selection This table shows the total number of completed M&A transactions with U.S. target and acquirer over the period from 1984 to 1997 initially obtained from the SDC/Thomson One Banker Deals database and the number of transactions included after the application of defined selection criteria (1) to (4) and after other exclusions. The sample selection criteria are: (1) The acquired block size is 5% to 49.9% and the acquirer does not have substantial (i.e. more than 5%) prior ownership in the target firm; (2) daily stock prices for the period t-220 to t+20 are available in the Datastream database and neither the acquirer nor the target have more than 75% zero-return days during the estimation period; (3) details of the block purchase reported in SDC database are confirmed by searching the Factiva database; (4) announcement of the block purchase does not take place in the middle of a tender offer process, subsequent to an announced takeover bid or with an expressly stated takeover intention. Other exclusions include concurrent material news announcements, the existence of prior ownership stakes (>4.9%) not reported in the SDC database, block purchases leading to de-facto majority control, transactions realized within one corporate entity, and double counts in the SDC.
Initial sample of U.S. minority block purchases between 1984 and 1997 Screened after criterion (1) Screened after criterion (2) Screened after criterion (3) Screened after criterion (4) After other exclusions - Final sample
Number
% of all reported transactions
731 451 204 198 175 141
100.0% 61.7% 27.9% 27.1% 23.9% 19.3%
Table 4–2: Distribution of Block Purchases by Year This table shows the distribution of the corporate block purchases by year. Year
Number of transactions
Percentage
Cumulative percentage
1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 Total
8 8 2 13 20 12 9 7 8 9 11 9 9 16 141
5.7% 5.7% 1.4% 9.2% 14.2% 8.5% 6.4% 5.0% 5.7% 6.4% 7.8% 6.4% 6.4% 11.3% 100.0%
5.7% 11.3% 12.8% 22.0% 36.2% 44.7% 51.1% 56.0% 61.7% 68.1% 75.9% 82.3% 88.7% 100.0% -
76
Table 4–3: Distribution of Block Purchases by Industry This table shows the distribution of block purchases by industry. Industry classification is based on historical primary SIC codes obtained from the Compustat database. Industry Natural resource extraction Real estate development Heavy construction Food products Apparel and textile mill products Lumber and wood products Furniture and fixtures Printing an publishing Chemicals and allied products Rubber and plastics Leather and footwear Stone, clay, and glass products Primary products and metal Industrial machinery Electronic/electric equipment Transportation equipment Instruments and related products Miscellaneous manufacturing Air transportation Communications Utilities and waste management Wholesale Retail Services Total
SIC Number of target Codes firms 10-14 5 15 2 16 1 20 8 22, 23 2 24 0 25 1 27 5 28 20 30 1 31 1 32 1 33, 34 10 35 14 36 18 37 1 38 5 39 1 45 2 48 8 49 6 50, 51 2 52-59 7 70-87 20 141
Proportion 3.5% 1.4% 0.7% 5.7% 1.4% 0.7% 3.5% 14.2% 0.7% 0.7% 0.7% 7.1% 9.9% 12.8% 0.7% 3.5% 0.7% 1.4% 5.7% 4.3% 1.4% 5.0% 14.2% 100%
Number of acquiring firms 8 0 0 6 2 1 3 4 12 2 0 1 2 11 14 1 18 2 2 7 5 3 4 33 141
Proportion 5.7% 4.3% 1.4% 0.7% 2.1% 2.8% 8.5% 1.4% 0.7% 1.4% 7.8% 9.9% 0.7% 12.8% 1.4% 1.4% 5.0% 3.5% 2.1% 2.8% 23.4% 100%
4.3.2 Measuring Corporate Relatedness 4.3.2.1
Benchmark Input-Output Accounts
Following Fan and Lang (2000) and Shahrur (2005), we use the benchmark input-output (IO) accounts for the U.S. economy provided by the Bureau of Economic Analysis at the U.S. Department of Commerce to construct a measure of corporate relatedness between the firms in our sample. The data source of these accounts are primarily economic censuses conducted by the U.S. Bureau of Census.38 In this study, we rely on data from the “Use table” of the benchmark accounts, which is a matrix containing the value of commodity flows between each pair of approximately 500 private sector, intermediate 38
For a detailed description of the benchmark IO tables, see Lawson (1997).
77
industries. More specifically, the Use table reports for each pair of industries, i and j, the dollar value of i’s output required to produce industry j’s total output, denoted as aij.
4.3.2.2
Method of Measuring Vertical Relatedness
Following Fan and Lang (2000) and Fan and Goyal (2006), we construct a measure of vertical relatedness based on the data from the Use table. In order to obtain a measure for forward vertical relatedness, we divide aij by the dollar value of industry j’s total output to get vij, which stands for the dollar value of industry i’s output necessary to produce one dollar’s worth of industry j’s output. Conversely, in order to capture a potential backward vertical relatedness between the firms, we divide aji by the dollar value of industry i’s output required to produce one dollar’s worth on industry j’s output. In the next step, we match a vertical relatedness coefficient to each transaction in our sample, according to the primary industry classification of the firms involved. Kahle and Walkling (1996) find that the Compustat industry classification is based on the current SIC code of a given firm, and hence does not account for the fact that many firms change their primary SIC code over time. As a consequence, we use Compustat’s historical SIC code data item, which represents the history of primary SIC codes for any particular firm.39 The information is available for all firms in our sample. In the third step, we have to convert the historical SIC codes to six-digit IO codes, as the Compustat database classifies industries only by their four-digit SIC code, whereas the Use table is based on a six-digit IO coding system. For this purpose, we use a conversion table constructed by Fan and Lang (2000) based on the conversion tables published by the Bureau of Economic Analysis. Their conversion table takes into account changes in industry definitions in the Use table over time. The first year for which the Use table is available is 1982. The data is published every five years by the Bureau of Economic Analysis. In order to measure the vertical relatedness of the firms in our sample, we use
39
Since Compustat reports the historical SIC primary SIC code only from 1987 onward, we use the 1987 historical SIC code for transactions in the years 1984, 1985 and 1986.
78
the 1982, 1987 and 1992 tables for block purchases that occur during the years 1984, 1985-1989, and 1990-1997, respectively.40 For each pair of firms involved in the block purchase, we then assign a vertical relatedness coefficient based on the primary IO industry classification of the firms. To illustrate this procedure and the construction of industry-level vertical relatedness coefficients, we provide an example from the plastics industry (see Table 4-4). Based on the 1992 Use table, the total output of the plastics industry (industry i) was $31,502 million (Qi), while the bags industry (industry j) had a total output of $8,389 million (Qj). The bags industry (industry j) used a total of $1,259 million in plastics (aij), while the plastics industry itself consumed $10 million of bags (aji) as input. In order to calculate the forward relatedness coefficient, we divide the amount of plastics used by the bags industry (aij = $1,259 million) by the total output of the bags industry (Qj = $8,389 million) to derive the forward vertical relation coefficient vij = 0.15. The value of vij can be interpreted as the dollar value of plastics consumed by the bags industry for each dollar of bags produced. In other words, the bag industry utilized $0.15 of plastics for each dollar of bags produced. Vice versa, the plastics industry only consumed $0.0003 (aji/Qi = 10/31,502) of bags for each dollar of plastics produced (vji). Using a 1% cutoff point to define significant relatedness (Shahrur (2005), Fan and Goyal (2006), Kale and Shahrur (2007)), we would classify this block purchase as forward vertically related, since there is a substantial forward relatedness from the perspective of the block purchaser (i.e. a plastics company).
40
As the SIC code based annual update of the 1992 benchmark tables is not publicly available, we use the vertical relatedness coefficients obtained from the 1992 benchmark table also for the 1995 to 1997 period.
Backward complementarity: Plastics and jth industries’ input flows correlation; corr(vki; vkj); k = 1…n, except i, j Adapted from Fan and Lang (2000)
Forward complementarity: Plastics and jth industries’ output flows correlation; corr(bik, bjk); k = 1…n, except i, j
Calculation of Complementarity coefficients
0.1128
0.0109
0.0032
Backward vertical relatedness: Value of j’s output used to produce $1 of plastics; vji = aji/Qi
10 31,502
Industry j’s output used by the plastics industry ($ millions); aji
Total plastics output ($ millions); Qi
0.1500
Forward vertical relatedness: Value of plastics used to produce $1 of j’s output; vij = aij/Qj
1,259 8,389
Total output of industry j ($ millions); Qj
2673, 2674
240702
Bags, except textile
Plastics used by industry j ($ millions); aij
Calculation of vertical relatedness coefficients
SIC code
Input-output code
Industry classification
Industry j
0.7262
0.4295
0.3589
31,502
11,143
0.0008
89,157
68
2800-2813, 2816, 2819, 2860-2869 (excl. 2861)
270100
Industrial inorganic and organic chemicals
0.9053
0.0130
0.0045
31,502
140
0.0096
9,736
93
2899
270406
Chemicals and chemical preparation
0.6272
0.2649
0.0002
31,502
7
0.0000
12,911
0
3200-3219, 3229-3239
350100
Glass and glass products
This table represents a quantitative example of the calculation of relatedness coefficients from the perspective of a company in the plastics industry. The data is based on the 1992 Use table provided by the Bureau of Economic Analysis.
An Illustration from the Plastics Industry
Table 4–4: Calculation of Vertical Relatedness and Complementarity Coefficients
79
80
However, as the plastics industry does not significantly rely on input from the bags industry (vji = 0.0003 < 0.01), backward vertical relatedness is not given in this example. In contrast, the industrial inorganic and organic chemicals industry contributes a large fraction of the inputs used by the plastics industry to produce 1$ worth of output (see Table 4.4), which is indicated by the high backward vertical relatedness coefficient of vji = 0.3589. Hence, we would classify a block purchase of a company from the plastics industry in a company from the industrial inorganic and organic chemicals industry as backward vertically related.
4.3.2.3
Method of Measuring Complementarity
In order to measure the degrees to which industries i and j share their input and output, we follow the methodology by Fan and Lang (2000). In the first step, for each industry the percentage of its output supplied to each intermediate industry k, denoted as bik, is calculated from the Use table. Then, a simple correlation coefficient between bik and bjk across all k except for i and j is calculated for each pair of industries i and j. A large correlation coefficient suggests a substantial overlap in the industries/markets to which industries i and j sell their products (Fan and Lang (2000)). In this case, forward complementarity is given. Similarly, we measure backward complementarity between the firms. For each pair of industries i and j a correlation coefficient across input industries (except i and j) between the input requirement coefficients vki and vkj of the two industries is derived. A large correlation coefficient indicates a high degree of backward complementarity, as the overlap in inputs required is high (Fan and Lang (2000)). Table 4-4 illustrates several examples of the construction of complementarity coefficients. For instance, the correlation of input flows between the plastics and chemicals and chemical preparation industry is 0.9053, which indicates a significant overlap in the inputs required by the two industries i and j and hence the presence of backward complementarity. Analyzing the forward complementarity coefficient between the plastics and the bags industry (0.0109), the low value is consistent with the intuitive
81
perception that these industries do not have a substantial overlap in the markets to which these industries sell their products. In this study, we use a cutoff point of 0.75 to define forward and backward complementarity.41 Table 4-5 presents descriptive statistics on the relatedness measures described above. Table 4–5: Sample Description This table contains descriptive statistics on the sample. Proportions are reported for binary variables. Mean and standard deviation are reported for continuous variables. A detailed description of the individual variables and the respective data sources are contained inTable 4-8. Panel A: General classification of block purchases N Horizontal Vertical /complementary Unrelated Total
Proportion of total sample 34.8% 24.8% 40.4% 100.0%
49 35 57 141
Proportion of non-horizontal sample 38.0% 62.0% 100.0%
Panel B: Descriptive statistics of relatedness variables Vertical relatedness coefficients Backward vertical relatedness Forward vertical relatedness Backward and/or forward vertical relatedness Complementarity coefficients Backward complementarity Forward complementarity Backward and/or forward complementarity
N
Proportion
Mean
Stdev.
14 16 24
15.2% 17.4% 26.1%
0.7% 1.2% -
2.2% 4.8% -
4 11 15
4.3% 12.0% 16.3%
33.0% 19.3% -
22.4% 27.3% -
N 141 141 141 74
Proportion 52.5%
Mean 13.0% -1.2 4.0% -
Stdev. 9.6% 1.1 81.8% -
Panel C: Descriptive statistics of control variables Block size Log. relative size Prior performance target Target high-tech industry
34.8% of the 141 transactions in our sample are horizontal transactions between firms which share the same six-digit IO code. This is slightly less than reported by Allen and Phillips (2000) who find that 53% of corporate block purchases involve firms in related industries (three-digit SIC level). 24.8% of all transactions are either vertically related and/or complementary. 40.4% of the block purchases are categorized as unrelated trans-
41
Other studies do not define exact cutoff points for complementarity. Hence, we repeated our analyses with alternative cutoff values of 0.25 and 0.50 and obtained similar results.
82
actions, i.e. transaction where the target’s and the acquirer’s industry are neither vertically integrated nor complementary.
4.3.3 Method of Measuring Abnormal Returns To assess the value implication of the corporate block purchases, we use standard event study methodology. Expected returns are generated from the market model parameters, estimated with daily returns from 220 days before the block purchase announcement (t-220) to 21 day before the announcement (t-21). Adjusted prices (taking into account dividend payments and changes to the capital structure) are used to calculate stock returns. As for the market return RM,t, we employ the Datastream U.S. non-financial total-return index. Abnormal returns are calculated as the difference between actual returns and estimated returns from the market model. The capital market reaction is evaluated through different samples, in which mean values aggregate the abnormal returns per each observation day (Henderson (1990)). We use five different event windows ranging between 2 days [-1;0] and 21 days [-10;10]. Following the suggestion by Harrington and Shrider (2007), we use the test statistic of Boehmer et al. (1991) to test the significance of cumulated abnormal returns. The test statistic z is used to account for the likely difference in cross-sectional return variance between the estimation period from t-220 to t-21 and the event window. The test statistic z follows a student t-distribution with T-2 degrees of freedom. The test results appear to be robust also in the absence of event-induced variance increases (Serra (2004)).
4.4
Results
4.4.1 Univariate Analysis Panel A of Table 4-6 presents the abnormal announcement period returns for the full sample of 141 corporate minority block purchases in the period 1984-1997. We find that all significant wealth gains in these transactions accrue to shareholders of target firms.
83
Table 4–6: Excess Returns Around Block Equity Purchases by Corporations This table shows the average cumulative abnormal returns for the total sample of 141 minority block purchases in panel A and for subsamples in panel B. Statistical significance at the 1%, 5%, and 10% level is denoted with ***, **, *. The statistical significance is tested using the test-statistic of Boehmer, Musumeci and Poulsen (1991) (z-statistic). Panel A: Total sample Target Event time period
CAAR
[-1;0] [-1;1] [-3;3] [-5;5] [-10;10] Panel B: Subsamples
6.33% 9.19% 10.18% 11.21% 13.55%
Acquirer
z-statistic 7.28*** 9.29*** 9.02*** 8.73*** 8.83***
Unrelated (N =57) CAAR z-statistic
CAAR 0.39% 0.50% 0.69% 0.83% 0.41%
Combined Entity
z-statistic 0.54*** 1.22*** 1.50*** 1.24*** 0.06***
Vert./Comp. (N=35) CAAR z-statistic
CAAR 0.77% 1.18% 1.66% 2.12% 2.86%
z-statistic 1.03*** 1.41*** 1.67*** 1.67*** 2.05***
Horizontal (N=49) CAAR z-statistic
[-1;0] event window Target Acquirer Combined entity
4.30% -0.17% -0.09%
3.87*** 0.72*** 0.75***
9.37% 0.81% 1.65%
5.09*** 0.37*** 1.09***
6.52% 0.74% 1.14%a
3.91*** 1.79*** 1.10***
[-1;1] event window Target Acquirer Combined entity
7.92% 0.22% 0.68%
4.89*** 0.71*** 2.03***
9.72% 0.42% 1.65%
5.89*** 0.95*** 1.09***
10.29% 0.88% 1.43%
5.81*** 1.98*** 2.08***
[-3;3] event window Target Acquirer Combined entity
9.38% -0.07% 0.61%
5.36*** 0.14*** 1.56***
11.04% 0.23% 2.18%
5.44*** 0.11*** 1.15***
10.49% 1.90% 2.52%
5.49*** 2.50*** 3.37***
[-5;5] event window Target Acquirer Combined entity
9.29% -0.18% 0.89%
5.41*** 0.42*** 1.14***
15.48% 1.09% 3.33%
5.57*** 0.54*** 1.21***
10.41% 1.80% 2.69%
4.35*** 1.80*** 2.51***
[-10;10] event window Target Acquirer Combined entity
13.06% -0.37% 2.55%
5.71*** 0.87*** 1.05***
15.31% -0.89% 2.66%
4.63*** 0.48*** 1.20***
12.87% 2.24% 3.36%
4.90*** 1.36*** 2.18***
Abnormal returns to block acquirers are low (between 0.39% and 0.83%) and statistically not significant. This is consistent with the observations of Allen and Phillips (2000), who attribute the low significance of abnormal acquirer returns to the large average discrepancy in firms size. We also analyze the abnormal returns that accrue to the combined entity by weighting the abnormal returns of each firm by its market capitalization at t-21.
84
We find that the combined abnormal returns are between 0.77% and 2.86% and statistically significant for three out of five event windows. With regard to the full sample, we find that target firms experience a highly significant average abnormal return (6.33% to 13.55%) upon the announcement of a block equity purchase by another corporation. This result is highly consistent with Allen and Phillips (2000) and Park et al. (2008) who report a cumulative abnormal return between 6.1% and 11.24% over the [-10;10] event window, respectively. As the capital markets may anticipate a subsequent takeover even in the absence of takeover related information upon a block purchase (Barclay and Holderness (1991), Sudarsanam (1996)), we also conduct an analysis where we exclude all transaction where the target company has been subject to a takeover within the one year period after the initial block purchase,42 leading to qualitatively the same results. In order to shed light on the univariate effect of corporate relatedness on abnormal returns, we analyze three subsamples presented in Panel B of Table 4-6. If the firms operate in the same six-digit IO industry, we classify these transactions as horizontal. We define transactions as vertically related/complementary if the firms do not operate in the same industry but are related either through vertical relatedness or complementarity. Finally, we define block purchases where target and acquirer operate in different six-digit IO industries and are neither vertically related nor complementary as unrelated transactions. Consistent with our initial hypotheses, abnormal returns to target firms, acquiring firms, and the combined entity are higher for vertically related/complementary and horizontal transactions for four out of five event windows. Conducting a standard mean-difference test for each of the pairs of subsamples (see Table 4-7), we find that the difference in combined entity returns of related (horizontal or vertical/complementary) vs. unrelated transactions and of vertical/complementary vs. unrelated transactions is statistically significant for the [-5;5] window. 42
19 target firms were taken over within the 1 year period after the block purchase, based on data we obtained from the Thomson ONE database.
85
Table 4–7: Mean Difference Test of Excess Returns This table shows the results of a standard mean difference test for specific pairs of subsamples. Statistical significance at the 1%, 5%, and 10% level is denoted with ***, **, *. CAAR [-1;1] Acquirer Combined entity Related (horizontal or vertical/complementary) vs. unrelated transactions Related 84 10.05%* 0.68%* 1.52%* Unrelated 57 7.92%* 0.22%* 0.68%* Mean difference 2.14%* 0.47%* 0.84%* t-value 1.087%* 0.591%* 1.077%* N
Horizontal vs. unrelated transactions Horizontal 49 Unrelated 57 Mean difference t-value -
Target
10.29%* 7.92%* 2.37%* 0.997%*
Target
CAAR [-5;5] Acquirer Combined entity
12.52%* 9.29%* 3.23%* 1.127%*
1.50%* -0.18%* 1.68%* 1.378%*
2.96%* 0.89%* 2.07%* 1.666*%
0.88%* 0.22%* 0.66%* 0.791%*
1.43%* 0.68%* 0.74%* 0.903%*
10.41%* 9.29%* 1.12%* 0.373%*
1.80%* -0.18%* 1.98%* 1.418%*
2.69%* 0.89%* 1.80%* 1.330%*
0.42%* 0.22%* 0.20%* 0.195%*
1.65%* 0.68%* 0.97%* 1.001%*
15.48%* 9.29%* 6.19%* 1,747*%
1.09%* -0.18%* 1.26%* 0.910%*
3.33%* 0.89%* 2.44%* 1,728*%
Horizontal vs. vertical/complementary transactions Horizontal 49 10.29%* 0.88%* Vertical/complementary 35 9.72%* 0.42%* Mean difference 0.56%* 0.46%* t-value 0.230%* 0.427%*
1.43%* 1.65%* -0.23%* -0.202%*
10.41%* 15.48%* -5.07%* -1.255%*
1.80%* 1.09%* 0.72%* 0.418%*
2.69%* 3.33%* -0.64%* -0.207%*
Vertical/complementary vs. unrelated transactions Vertical/complementary 35 9.72%* Unrelated 57 7.92%* 1.81%* Mean difference t-value 0.767%*
This finding indicates that the capital market positively values the potential benefits that stem from strategically related transactions relative to purely unrelated transactions. Fan and Goyal (2006) provide similar evidence for a sample of 2,162 mergers during 1962 to 1996. They find that the combined wealth effect of vertical mergers is statistically significant at 2.5% over the [-1;1] for vertical mergers compared with 1.9% for the entire sample. They also report that the difference between the wealth effects for vertical mergers and those for all mergers is significant at the 1% level. Since our results may be due to cross-sectional differences across the transactions, we further extend our analysis and conduct multivariate analyses in the following section.
86
4.4.2 Cross-Sectional Regression Analysis Following Fan and Goyal (2006), we focus our analysis on the combined wealth effect of the transaction, since the sum of economic benefits from a relationship should accrue to both, the target and the acquirer. We estimate regressions using the cumulative abnormal combined entity return over the [-1;1] and the [-5;5] event window as dependent variable. 43 Table 4-8 provides an overview over independent variables used. Table 4-9 reports the regression specifications for the cumulative abnormal returns to the combined entity. All test statistics are computed using White’s heteroskedasticity-consistent covariance matrix (White (1980)). We do not find evidence for the existence of multicollinearity. The variables included in our models are structured following our initial hypotheses. We design three different regression models that give an increasingly detailed view on the type of relatedness. In model 1, we only include the broad dummy variable “related” in order to broadly account for relatedness. In model 2, we further differentiate between vertical relatedness and complementarity. Finally, model 3 also includes dummy variables accounting for the specific direction of vertical relatedness and complementarity (forward vs. backward). In order to account for other factor that may influence the regression results, we include four control variables. First, we include into our sample the proportion of target stock owned after the block purchase as a measure of block size, as the likelihood that the benefit of monitoring exceeds the cost increases with the size of the blockholding (Park et al. (2008)). In addition, blockholders owning large stakes are more likely to monitor management actions since they face a liquidity problem (Maug (1998)).
43
Using the cumulative abnormal returns of other event windows as the dependent variable leads to similar qualitative results.
87
Table 4–8: Variable and Source Description This table contains the definitions and data sources for each variable collected. Variable name
Definition
Sources
Type of corporate relatedness Horizontal
Dummy variable that equals one if the acquirer and the target share the same six-digit IO code
Vertical/ complementary
Dummy variable that equals one if the transaction is nonhorizontal and if either vertical relatedness and/or complementarity is given Vertical relatedness Dummy variable that equals one if the transaction is nonhorizontal and if either forward and/or vertical relatedness is given Forward vertical rela- Dummy variable that equals one if the transaction is nontion horizontal and if the forward vertical relatedness coefficient vij is larger than 1% Backward vertical re- Dummy variable that equals one if the transaction is nonlation horizontal and if the backward vertical relatedness coefficient vji is larger than 1% Complementarity Dummy variable that equals one if the transaction is nonhorizontal and if either forward and/or backward complementarity is given Forward complemen- Dummy variable that equals one if the transaction is nontarity horizontal and if the forward complementarity coefficient corr(bik, bjk) is larger than 75% Backward compleDummy variable that equals one if the transaction is nonmentarity horizontal and if the forward complementarity coefficient corr(vik, vjk) is larger than 75% Control variables Block size Proportion of the common owned after the block purchase Prior performance tar- The buy-and-hold return of the target relative to the buy-andget hold return of the specific country market index over the estimation period Log relative size The natural logarithm of the market value of the target divided by the acquirer’s market value at t-21 High-tech industry Dummy variable taking the value of one if the target company operates in a high-technology industry
Bureau of Economic Analysis, COMPUSTAT Bureau of Economic Analysis, COMPUSTAT Bureau of Economic Analysis, COMPUSTAT Bureau of Economic Analysis, COMPUSTAT Bureau of Economic Analysis, COMPUSTAT Bureau of Economic Analysis, COMPUSTAT Bureau of Economic Analysis, COMPUSTAT Bureau of Economic Analysis, COMPUSTAT SDC/ Thomson One Banker, Factiva Datastream
Datastream Hecker (1999)
88
Table 4–9: Determinants of Cumulative Average Abnormal Combined Entity Returns based on Primary Industry Classification This table shows estimation results for OLS-regression models with the cumulative abnormal return for the [-3;3] and [-5;5] interval as the dependent variables. The variables and data sources are laid out in Table 4-8. T-values are in brackets. Statistical significance at the 1%, 5%, and the 10% level is denoted with ***, **, *. All test statistics are computed using a heteroskedasticity-consistent covariance matrix (White (1980)). No evidence for multicollinearity was found. Dep. variable: CAAR [-1;1]
Dep. variable: CAAR [-5;5]
Model 1
Model 2
Model 3
Model 4
Model 5
Model 6
Intercept
0.024)*** (2.793)***
0.024)*** (2.851)***
0.022)*** (2.523)***
0.032)*** (2.409)***
0.033)*** (2.478)***
0.031)*** (2.278)***
Horizontal
0.007)*** (0.893)***
0.007)*** (0.854)***
0.009)*** (1.173)***
0.023)*** (1.648)***
0.021)*** (1.510)***
0.024)*** (1.725)***
Vertical/complementary
0.008)*** (0.899)***
0.023)*** (1.592)***
Measures of vertical integration Vertical relatedness
0.002)*** (0.238)***
0.000)*** (0.041)***
Forward vertical relatedness
0.021)*** (1.514)***
0.031)*** (1.511)***
Backward vertical relatedness
0.001)*** (0.091)***
-0.007)*** (-0.32)***
Measures of complementarity Complementarity
0.014)*** (1.055)***
Forward complementarity
0.044)*** (1.940)*** -0.020)*** (-1.256)***
0.016)*** (0.468)***
0.027)*** (1.827)***
0.056)*** (2.100)***
Backward complementarity Control variables Block size
-0.001)*** (-0.560)***
-0.000)*** (-0.562)***
0.013)*** (4.041)***
0.012)*** (3.849)***
Prior performance target
-0.016)*** (-5.195)***
-0.016)*** (-5.16)***
-0.016)*** (-4.849)***
High-tech industry
-0.002)*** (-0.310)***
-0.003)*** (-0.346)***
-0.003)*** (-0.41)***
Log relative size
N Adj. R2
-0.000)*** (-0.514)***
-0.008)*** (-1.567)***
-0.001)*** (-1.532)***
-0.000)*** (-1.545)***
0.012)*** 0.020)*** (3.914)*** )***(4.354)***
0.018)*** (3.970)***
0.019)*** (3.985)***
-0.010)*** (-1.193)***
-0.010)*** (-1.216)***
-0.011)*** (-1.185)***
0.010)*** (1.040)***
0.012)*** (0.953)***
0.010)*** (0.780)***
141
141
141
141
141
141
17.7%
17.5%
20.5%
9.6%
10.7%
11.4%
Second, we include relative size as the logarithm of the ratio of the market capitalization of the target relative to that of the bidder, with both values being measured 20 days prior to the announcement of the takeover. Similar approaches are used by Servaes (1991), Mulherin and Boone (2000), and Shahrur (2005).
89
Third, the prior performance of the target is measured as the buy-and-hold return of the target firms’s stock relative to the buy-and-hold index return over the estimation period. As management quality and stock price performance are found to be related (e.g. Warner et al. (1988), Denis and Kruse (2000)), larger potential benefits through additional monitoring from a new outside blockholder should be expected if the target company is performing poorly. Consequently, we expect an inverse relation between abnormal returns and prior performance. Finally, we also include a dummy variable that equals the value of 1 if the target company operates in a high-tech industry. Given the inherent potential for information asymmetries, contracting problems between corporations seem particularly likely in high-technology industries. Pisano (1989) argues that partial ownership may be particularly effective for activities such as R&D that are costly to govern through contracts but are subject to incentive losses when internalized completely. This view is supported by Fee et al. (2006) who find that equity stakes are much more common in suppliercustomer relationships where the supplier is a R&D intensive firm. Accordingly, Allen and Phillips (2000) find that targets operating in industries with high R&D expenses show significant improvements in operating cash-flows and increases in investment expenditures following corporate equity block purchases, which they partly contribute to the mitigation of information problems between the corporations involved in the block purchase. In order to objectively evaluate the R&D-intensity of the target firm, we use the classification scheme of Hecker (1999), who systematically identifies 31 three-digit SIC codes that comprise high-technology industries based on measures of industry employment in both R&D and technology oriented occupations.44 We first analyze the effect of relatedness on abnormal returns using our broad regression specification (models 1 and 4) that does not account for the specific type of relatedness, but only uses a dummy variable that equals one if any form of vertical relatedness or complementarity is given. We find both horizontal block purchases and related 44
Hecker (1999) classifies those industries as high-technology industries where the number of R&D workers and technology oriented occupations accounts for a proportion of employment that is at least twice the average of all industries included in the analysis.
90
block purchases to have a positive effect on the capital market’s valuation of the transaction, consistent with our initial hypotheses. However, this relationship is not statistically significant. We then use the more detailed models (2, 3, 5, 6) to proxy the type of corporate relatedness between the block purchaser and the target more specifically. Consistent with our observations in the univariate analysis and the regression results using the broad specification of relatedness, we find the influence of the “horizontal” dummy variable to be positive. However, we find the coefficient to be statistically significant in model 6 only. We cannot observe a statistically significant relationship between the dummy variables that proxy vertical relatedness and the observed abnormal returns. The sign of the coefficients is positive as initially hypothesized with only one exception. In contrast, we find the dummy variables “complementarity”, and “backward complementarity” to be statistically significant and positive in models 3, 5 and 6. With regard to the control variables, we find a highly significant relationship between both “log relative size” and “prior performance” and the excess return to the combined entity. We assume that positive impact of the relative size of the target to that of the bidder is due to the relatively larger exposure the acquirer may have to the performance of the target company resulting in an increased incentive to effectively monitor the target. Moreover, one may contribute this finding to a larger potential for inter-company strategic collaboration if both companies have a critical size. Consistent with our initial hypothesis, transactions where the target performed poorly prior to the block purchase experience significantly higher abnormal returns (models 1-3). This finding is consistent with Akhigbe et al. (2004) who report a significant negative relationship between the announcement returns of block purchases and prior stock performance. As the stock price performance is frequently related to management quality, a possible explanation for this finding is that the capital market expects a performance enhancement through more effective monitoring by a new blockholder.
91
Summing up, our results suggest that the relatedness of the block acquirer and the target firm has a positive wealth effect, consistent with our initial hypotheses. However, given the low statistical significance of these findings, it is questionable whether this effect is substantial. As the highly significant influence of our control variables suggests, other factors such as the potential presence of monitoring or agency problems seem to explain a significantly larger fraction of the observed abnormal returns.
4.4.3 The Effect of Secondary Segments Although the focus of this paper lies on the relationship of the primary industry segments of firms, we also analyze the empirical implications of multisegment firms in order to guarantee the robustness of our findings. Maksimovic and Philipps (2001) show that many mergers are between primary and secondary divisions of firms. Fan and Goyal (2006) find greater wealth creation of vertical mergers on both the primary and secondary segment level. A pure focus on the primary segments of the firms raises two potential concerns with regard to the robustness of empirical tests. First, a block purchase that is classified as vertical on the basis of the primary segment of the firm may have additional horizontal relationships with secondary segments. Second, the analysis of relatedness between primary segments may not capture relatedness at the secondary level (Fan and Goyal (2006)). Consequently, we address these concerns by analyzing relatedness on both the primary and secondary segment level. For this purpose, we obtain historical secondary segment information for each firm in our sample. We derive 112 transactions for which we can find consistent information for both companies on the Compustat segment database. For each year, we rank the individual segments of each firm by their reported sales. We then obtain the name and the historical four-digit SIC code for the two largest segments. We then assign a six-digit IO code to each segment, using the conversion table constructed by Fan and Lang (2000). If the firms share the same six-digit IO code on either the primary or the secondary segment level, we classify the transaction as horizontal. For the non-horizontal transactions, we
92
calculate individual vertical relatedness and complementarity coefficients for each intersegment relationship, including both primary and secondary segments. The vertical relatedness and complementarity dummies take a value of one if the vertical relatedness coefficients and the complementarity coefficients of any pair of segments exceed the cutoff points of 1% and 75%, respectively. Table 4-10 reports descriptive statistics on the control sample of 112 transactions. Table 4–10: Descriptive Statistics of Control Sample This table contains descriptive statistics on the sample. We classify a transaction a horizontal if the firms share the same six-digit IO code for at least one of the two largest divisions. We classify a transaction as related if we find vertical relatedness and/or complementarity between the two largest divisions of the firms. If do not find any relatedness, we classify the transaction as unrelated. The divisional data is obtained from the Compustat segment database. The four-digit segment SIC codes are converted into sixdigit IO codes using the conversion table provided by Fan and Lang (2000). Panel A: General classification of block purchases Classification based industry classi- Classification based on primary fication including secondary industry classification segments N Proportion N Proportion Horizontal 41 36.6% 40 35.7% Vertical/complimentary 41 36.6% 30 26.8% Unrelated 30 26.8% 42 37.5% Total 112 100.0% 112 100.0%
Panel B: Segment information
Number of companies with one segment Number of companies with two or more segments Total Average size of secondary segment relative to primary segment
N 88 24
Target firms Proportion 78.6% 21.4%
112
100.0% 42.4%
Acquiring firms N Proportion 44 39.3% 68 60.7% 112
100.0% 46.1%
As supposed, the inclusion of secondary segment information leads to a larger proportion of related and horizontal transactions (73.2%) relative to our initial classification based on primary segment information only (62.5%). We find that 50.0% of the transactions include at least one company with more than one business segment. The percentage of multisegment acquiring firms (60.7%) is higher than for target firms (21.4%), which we attribute to the large average difference in firm size. The relative size of the secon-
93
dary segment’s revenue to that of the primary segment is 42.4% for target companies and 46.1% for acquiring companies. Table 4–11: Excess Returns around Block Purchases for Control Sample This table shows the average cumulative abnormal returns for the control sample of 121 minority block purchases in panel A and for subsamples in panel B. Statistical significance at the 1%, 5%, and 10% level is denoted with ***, **, *. The statistical significance is tested using the test-statistic of Boehmer, Musumeci and Poulsen (1991) (z-statistic). Panel A: Total sample Target Event time period
CAAR
[-1;0] [-1;1] [-3;3] [-5;5] [-10;10]
6.52% 8.74% 9.23% 10.79% 13.35%
Acquirer
z-statistic 6.89*** 7.89*** 8.00*** 7.78*** 7.86***
CAAR 0.38% 0.59% 0.47% 0.59% 0.37%
z-statistic 0.34*** 1.05*** 0.86*** 0.84*** 0.06***
Combined Entity CAAR 0.78% 1.17% 1.26% 1.68% 2.58%
z-statistic 1.12*** 1.25*** 1.35*** 1.29*** 1.57***
Panel B: Subsamples Unrelated (N =30) CAAR z-statistic
Vert./comp. (N=41) CAAR z-statistic
Horizontal (N=41) CAAR z-statistic
[-1;0] event window Target Acquirer Combined entity
4.92% -0.12% 0.15%
3.41*** 1.38*** 0.00***
9.28% 0.27% 0.87%
5.43*** 0.16*** 1.46***
4.93% 0.85% 1.16%
3.15*** 1.81*** 1.03***
[-1;1] event window Target Acquirer Combined entity
7.75% 0.27% 0.95%
3.83*** 0.14*** 1.25***
10.01% 0.30% 0.81%
5.59*** 0.47*** 1.10***
8.18% 1.13% 1.69%
4.91*** 2.09*** 1.16***
[-3;3] event window Target Acquirer Combined entity
9.00% -0.08% 0.33%
3.71*** 0.36*** 0.30***
10.70% -0.63% 0.62%
5.78*** 0.62*** 0.79***
7.93% 1.96% 2.59%
4.30*** 2.05*** 1.28***
[-5;5] event window Target Acquirer Combined entity
8.72% 0.32% 1.04%
4.12*** 0.47*** 1.34***
14.21% -0.36% 0.99%
5.17*** 0.49*** 0.35***
8.87% 1.75% 2.83%
4.31*** 1.34*** 1.22***
12.00% 1.07% 4.02%
4.10*** 0.37*** 1.65***
15.32% -2.14% 0.17%
4.63*** 1.54*** 0.72***
12.38% 2.37% 3.92%
4.87*** 0.99*** 1.50***
[-10;10] event window Target Acquirer Combined entity
Table 4-11 shows the excess returns for both the full sample and subsamples over different event windows. In comparison to the results of our initial sample reported in Table 4-6, we still observer larger excess returns for related transactions than for unre-
94
lated transactions in most cases. These differences are, however, not statistically significant (see Table 4-12). Table 4–12: Mean Difference Test of Excess Returns for Control Sample This table shows the results of a standard mean difference test for specific pairs of subsamples. Statistical significance at the 1%, 5%, and 10% level is denoted with ***, **, *. CAAR [-1;1] Acquirer Combined Entity Related (horizontal or vertical/complementary) vs. unrelated transactions Related 82 9.10% 0.71% 1.25% Unrelated 30 7.75% 0.27% 0.95% Mean difference 1.35% 0.45% 0.30% t-value 0.598% 0.441% 0.321% N
Horizontal vs. unrelated transactions Horizontal 41 Unrelated 30 Mean difference t-value -
Target
8.18% 7.75% 0.43% 0.170%
Target
CAAR [-5;5] Acquirer Combined Entity
11.54% 8.72% 2.82% 0.814%
0.69% 0.32% 0.37% 0.247%
1.91% 1.04% 0.87% 0.585%
1.13% 0.27% 0.86% 0.820%
1.69% 0.95% 0.74% 0.750%
8.87% 8.72% 0.15% 0.049%
1.75% 0.32% 1.43% 0.788%
2.83% 1.04% 1.80% %1.094%
0.30% 0.27% 0.03% 0.025%
0.81% 0.95% -0.14% -0.127%
14.21% 8.72% 5.49% 1.286%
-0.36% 0.32% -0.68% -0.425%
0.99% 1.04% -0.05% -0.031%
Horizontal vs. vertical/complementary transactions Horizontal 41 8.18% 1.13% Vertical/complementary 41 10.01% 0.30% Mean difference -1.83% 0.83% t-value -0.800% 0.791%
1.69% 0.81% 0.88% 0.911%
8.87% 14.21% -5.34% -1.397%
1.75% -0.36% 2.11% 1.371%
2.83% 0.99% 1.85% 1.154%
Vertical/complementary vs. unrelated transactions Vertical/complementary 41 10.01% Unrelated 30 7.75% Mean difference 2.26% t-value 0.868%
Similarly, our regression analysis (Table 4-13) does not reflect this observation. Although the coefficients of the dummy variables “horizontal” and “related” are positive for both model 1 and 4, we do not find them to be statistically significant. We make a similar observation for the detailed dummy variables in models 2, 3, 5 and 6.
95
Table 4–13: Determinants of Cumulative Average Abnormal Combined Entity Returns for Control Sample This table shows estimation results for OLS-regression models with the cumulative abnormal return for the [-3;3] and [-5;5] interval as the dependent variables. The variables and data sources are laid out in Table 4-8. T-values are in brackets. Statistical significance at the 1%, 5%, and the 10% level is denoted with ***, **, *. All test statistics are computed using a heteroskedasticity-consistent covariance matrix (White (1980)). No evidence for multicollinearity was found. Dep. variable: CAAR [-1;1] Model 1 Model 2 Model 3 0.025)*** 0.025)*** 0.020)*** (2.027)*** (2.177)*** (1.840)***
Dep. variable: CAAR [-5;5] Model 4 Model 5 Model 6 0.009)*** 0.032)*** 0.025)*** (0.453)*** (1.838)*** (1.504)***
Horizontal
0.005)*** (0.597)***
0.019)*** (1.217)***
Vertical/complimentary
0.000)*** (0.033)***
Intercept
0.005)*** (0.611)***
0.009)*** (1.013)***
0.017)*** (1.114)***
0.023)*** (1.495)***
0.006)*** (0.413)***
Measures of vertical integration Vertical relatedness
0.002)*** (0.248)***
-0.009)*** (-0.017)***
Forward vertical relatedness
0.007)*** (0.763)***
0.006)*** (0.420)***
Backward vertical relatedness
0.012)*** (0.909)***
0.020)*** (0.847)***
Measures of complementarity Complementarity
-0.006)*** (-0.374)***
Forward complementarity
-0.003)*** (-0.131)*** -0.021)*** (-1.429)***
-0.024)*** (-1.486)***
0.002)*** (0.079)***
0.007)*** (0.169)***
-0.000)*** -0.000)*** -0.000)*** (-0.292)*** (-0.260)*** (-0.013)***
0.000)*** -0.001)*** -0.000)*** (0.701)*** (-0.878)*** (-0.665)***
Backward complementarity Control variables Block size Log relative size Prior performance target High-tech industry N Adj. R2
0.012)*** (3.277)***
0.013)*** (3.386)***
0.013)*** (3.457)***
0.017)*** (3.137)***
0.014)*** (2.628)***
0.015)*** (2.724)***
-0.014)*** -0.014)*** -0.014)*** -0.021)*** -0.008)*** -0.008)*** (-6.965)*** (-7.093)*** (-6.918)*** (-6.370)*** (-1.155)*** (-1.340)*** 0.001)*** (0.178)***
0.001)*** (0.203)***
0.001)*** (0.193)***
0.013)*** (1.019)***
0.006)*** (0.417)***
0.006)*** (0.425)***
112
112
112
112
112
112
16.1%
15.5%
16.6%
3.6%
2.6%
2.8%
In contrast to our initial regression analyses, none of the relatedness proxies has a statistically significant impact on the capital market’s valuation of a block purchase. Given the fact that the relatedness dummy variables in our initial sample are based on the primary industry classification code and hence represent the key industries the target and the acquirer operate in, the lower significance of the relatedness dummy variables in our
96
control sample may be explained with the lower economic relevance of strategic relationships between subordinated (i.e. smaller) operating industries. With regard to the control variables, we find both the “log relative size” and “prior performance” variable to be statistically significant, which is consistent with our results of our primary industry classification sample. We regard this as additional confirmation of the hypothesis that the resolution of agency problems may be an important source for wealth gains in minority block purchases between corporations.
4.5
Conclusion
In this study, we investigate the share price reactions to corporate minority block purchases for a sample of 141 U.S. block purchases between 1984 and 1997. We argue that the type and degree of relatedness between the firms in our sample should have a significant effect on the observed market reaction, due to potential benefits from higher productive efficiency and increased market power. For this reason, we calculate different measures of relatedness following our main hypotheses with regard to the potential sources of wealth creation. We then ask whether these have a differential impact on abnormal returns upon the announcement of the block purchase. In the first part of our analysis, we document a significant wealth creation for target companies. The finding of significant positive announcement returns is consistent with prior empirical evidence (e.g. Mikkelson and Partch (1986), Wruck (1989), Allen and Phillips (2000)). In line with Allen and Phillips (2000), we also report significant positive value-weighted combined entity abnormal returns. This finding strengthens the assumption that the formation of corporate minority ownership blocks has specific potential benefits that translate into a positive wealth effect. Differentiating between horizontal, vertically related/complementary and unrelated block purchases, we find that related and horizontal block purchases lead to higher excess returns for targets, acquirers and the combined entity in most cases. This observation is also consistent with Fan and Goyal (2006) who report significantly higher excess returns for vertically related mergers.
97
Controlling for other factors such as R&D-intensity and relative size in cross-section, we can still find a positive relationship between the value-weighted combined abnormal returns and our proxy variables for horizontal relatedness, vertical relatedness, and complementarity. Although we find this relationship to be statistically significant only in a limited number of cases, we interpret these results as an indication for the potential benefits that stem from corporate block ownership of related relative to unrelated corporations. These results are similar when we expand our analysis to secondary segment information in order to account for relatedness beyond the primary industry classification. In conclusion, our findings suggest a positive effect of industry relatedness on the capital market’s valuation of corporate block purchases. We attribute this finding to the higher potential for efficiency gains and increased market power if the firms operate in related industries. However, the statistical significance of this finding is relatively low. Validations of this interpretation in further detail should be a fruitful topic for future research. For example, one could account the characteristics of established customersupplier relationships using data from the Compustat industry segment files based on a similar methodology to that devised by Fee and Thomas (2004) and Fee et al. (2006).
99
5
Concluding Remarks
This doctoral thesis follows the research objective of improving our understanding of the motives and consequences of minority equity purchases. It focuses on the wealth effects of minority block purchases and addresses three empirical research gaps in this context. First, we explore the implications of minority block purchases in German exchangelisted companies in chapter 2. Our analysis builds on the existing empirical evidence from U.S. block purchases and expands the analysis to a bank-based system of corporate governance characterized by a substantially higher ownership concentration and a considerable power associated with minority stakes. Analyzing a new set of data with 85 block purchases with German targets between 1997 and 2006, we can document a significant positive wealth effect for target companies. Our findings are consistent with the U.S. evidence, despite the structural differences in the corporate governance systems. The effect of the existing ownership structure of the target prior to the transaction is statistically neglectable. Differentiating between activist, strategic and financial acquirers, we find that block purchases by activist investors are accompanied by significantly higher abnormal returns than those of purely financial investors. We attribute this finding to either the market’s perception on superior monitoring or stock picking ability by activist investors such as private equity funds. Overall, we believe that this finding underlines the importance of taking into account blockholder heterogeneity in empirical studies on blockholders and corporate governance. Expanding on the notion of blockholder heterogeneity, we then focus on nonfinancial corporations as blockholders in chapter 3. We can show that there is only fragmented and incomplete empirical evidence on the wealth effects of corporate block purchases, despite the specific characteristics and theoretical benefits of corporate (minority) equity ownership. Analyzing a sample of 113 European minority equity purchases by non-financial corporations during the 1993 to 2006 period, we find that corporate block purchases lead to significant positive abnormal returns for target companies. More importantly, we can document that the combined value-weighted returns are positive and
100
significant, which indicates that corporate equity ownership is perceived by the capital markets as value creating in total. Moreover, we can exclude systematic wealth transfers from the purchasing firms to the target firms and vice versa. We cannot, however, confirm our initial hypotheses with regard to the factors that relate to specific benefits of corporate block ownership such as the alignment of incentives or the mitigation of information problems in business relationships. We believe that this finding may be either due to the limited economic magnitude of these benefits or to an incomplete or lagged valuation by the capital markets. In contrast, we can show that the stock price reaction is significantly more positive for both target firms and acquiring firms if the target shows symptoms of ineffective monitoring or existing agency problems. This finding may indicate that the potential benefits from a new outside blockholder are much more apparent if the target is performing poorly, especially when compared to more subtle, operational benefits in existing business relationships. Given the large research gap with regard to corporate blockholders, we further concentrate on specific role of strategic relatedness in the context of corporate block purchases (chapter 4). Based on numerous theoretical considerations that lead to the hypothesis that potential benefits of minority block purchases depend on the type of corporate relatedness between the acquirer and the target, we can identify an empirical research gap. Applying a comprehensive methodology to measure the type and degree of corporate relatedness, we analyze a sample of 141 U.S. minority block purchases during the 1984 to 1997 period. We find that a significant portion of corporate block purchases (59.6%) is between related companies. Consistent with our results in chapter 2, we can document significant positive value-weighted combined entity returns. With regard to the effect of corporate relatedness on excess returns, we can find a positive relationship between the value-weighted abnormal returns and our empirical measures of corporate relatedness. Although we find this relationship to be statistically significant only in a limited number of cases, we interpret these results as an indication of the potential benefits related to corporate relatedness. In sum, our analyses consistently show that the theoretical notion of blockholder heterogeneity has as measurable empirical effect. The capital markets seem to take into
101
account the specific preferences and skills of large blockholders in the valuation of changes in the ownership structure of corporations. Remarkably, we can show that even on the level of specific types of blockholders (e.g. corporate blockholders), there is a significant influence of highly specific factors such as corporate relatedness. We believe that further empirical research in this area should take into account the high degree of specificity with regard to blockholder identity. In addition, our analyses show that the wealth effects from corporate block purchases are not only determined by factors relating to the new blockholder, but also by attributes and characteristics of the target company (e.g. existing agency problems). A deeper understanding of the interaction effect between blockholder characteristics on the one hand and target characteristics on the other hand should be regarded as another motivation for further research in this area.
103
References Aboody, David, and Baruch Lev, 2000, Information Asymmetry, R&D, and Insider Gains, Journal of Finance 55, 2747-2766. Aghion, Phillipe, and Jean Tirole, 1994, The Management of Innovation, The Quarterly Journal of Economics 109, 1185-1209. Agrawal, Anup, and Gershon N. Mandelker, 1990, Large Shareholders and the Monitoring of Managers: The Case of Antitakeover Charter Amendments, Journal of Financial & Quantitative Analysis 25, 143-161. Akhigbe, Aigbe, Jeff Madura, and Carolyn Spencer, 2004, Partial Acquisitions, Corporate Control, and Performance, Applied Financial Economics 14, 847-857. Akhigbe, Aigbe, Anna D. Martin, and Ann Marie Whyte, 2007, Partial Acquisitions, the Acquisition Probability Hypothesis, and the Abnormal Returns to Partial Targets, Journal of Banking & Finance 31, 3080-3101. Allen, Jeffrey W., and Gordon M. Phillips, 2000, Corporate Equity Ownership, Strategic Alliances, and Product Market Relationships, Journal of Finance 55, 2791-2815. Banerjee, Saugata, Benoit Leleux, and Theo Vermaelen, 1997, Large Shareholdings and Corporate Control: An Analysis of Stake Purchases by French Holding Companies, European Financial Management 3, 23-43. Barclay, Michael J., and Clifford G. Holderness, 1991, Negotiated Block Trades and Corporate Control, Journal of Finance 46, 861-878. Barclay, Michael J., Clifford G. Holderness, and Dennis P. Sheehan, 2007, Private Placements and Managerial Entrenchment, Journal of Corporate Finance 13, 461484. Barth, Mary E., Ron Kasznik, and Maureen F. McNichols, 2001, Analyst Coverage and Intangible Assets, Journal of Accounting Research 39, 1-34. Becht, Marco, and Ekkehart Boehmer, 2003, Voting Control in German Corporations, International Review of Law & Economics 23, 1-29. Beiner, Stefan, Wolfgang Drobetz, Markus M. Schmid, and Heinz Zimmermann, 2006, An Integrated Framework of Corporate Governance and Firm Valuation, European Financial Management 12, 249-283. Beitel, Patrick, Dirk Schiereck, and Mark Wahrenburg, 2004, Explaining M&A Success in European Banks, European Financial Management 10, 109-139. Berle, Adolf A., and Gardner C. Means, 1932, The Modern Corporation and Private Property, New York: McMillian, 1932. Bethel, Jennifer E., Julia Porter Liebeskind, and Tim Opler, 1998, Block Share Purchases and Corporate Performance, Journal of Finance 53, 605-634.
104
Boehmer, Ekkehart, Jim Musumeci, and Annette B. Poulsen, 1991, Event-Study Methodology Under Conditions of Event-Induced Variance, Journal of Financial Economics 30, 253-272. Brav, Alon, Wei Jian, Frank Partnoy, and Randall S. Thomas, 2008, The Returns to Hedge Fund Activism, ECGI - Law Working Paper No. 098/2008. Campa, José Manuel, and Ignacio Hernando, 2004, Shareholder Value Creation in European M&As, European Financial Management 10, 47-81. Chan, Su Han, John W. Kensinger, Arthur J. Keown, and John D. Martin, 1997, Do strategic alliances create value?, Journal of Financial Economics 46, 199-221. Chen, Yongmin, 2001, On Vertical Mergers and Their Competitive Effects, The RAND Journal of Economics 32, 667-685. Choi, Dosoung, 1991, Toehold Acquisitions, Shareholder Wealth, and the Market for Corporate Control, Journal of Financial & Quantitative Analysis 26, 391-407. Coase, Ronald H., 1937, The Nature of the Firm, Economica 4, 386-405. Croci, Ettore, 2007, Corporate Raiders, Performance and Governance in Europe, European Financial Management 13, 949-978. Cronqvist, Henrik, and Rüdiger Fahlenbrach, 2007, Large Shareholders and Corporate Policies, The Review of Financial Studies Forthcoming. Demsetz, H., and K. Lehn, 1985, The Structure of Corporate Ownership: Causes and Consequences, Journal of Political Economy 93, 1115-1177. Denis, David J., Diane K. Denis, and Keven Yost, 2002, Global Diversification, Industrial Diversification, and Firm Value, Journal of Finance 57, 1951-1979. Denis, David J., and Timothy A. Kruse, 2000, Managerial Discipline and Corporate Restructuring Following Performance Declines, Journal of Financial Economics 55, 391-424. Dherment-Ferere, Isabelle, Jens Köke, and Luc Renneboog, 2004, Corporate Monitoring by Blockholders in Europe: Empirical Evidence of Managerial Disciplining in Belgium, France, Germany and the UK, ZEW discussion paper No. 01-24. Dyck, Alexander, and Luigi Zingales, 2004, Private Benefits of Control: An International Comparison, Journal of Finance 59, 537-600. Ezrachi, Ariel, and David Gilo, 2006, EC Competition Law and the Regulation of Passive Investments Among Competitors, Oxford Journal of Legal Studies 26, 327349. Faccio, Mara, and Larry H. P. Lang, 2002, The Ultimate Ownership of Western European Corporations, Journal of Financial Economics 65, 365-395. Faccio, Mara, and David Stolin, 2006, Expropriation vs. Proportional Sharing in Corporate Acquisitions, Journal of Business 79, 1413-1444.
105
Fan, Joseph P.H., and Vidhan Goyal, K. Goyal, 2006, On the Patterns and Wealth Effects of Vertical Mergers, Journal of Business 79, 877-902. Fan, Joseph P.H., and Larry H. P. Lang, 2000, The Measurement of Vertical Relatedness: An Application to Vertical Diversification, Journal of Business 73, 629-660. Fee, C. Edward, Charles J. Hadlock, and Shawn Thomas, 2006, Corporate Equity Ownership and the Governance of Product Market Relationships, Journal of Finance 61, 1217-1251. Fee, C. Edward, and Shawn Thomas, 2004, Sources of Gains in Horizontal Mergers: Evidence from Customer, Supplier, and Rival Firms, Journal of Financial Economics 74, 423-460. Ferguson, Michael F., 1994, Ownership Structure, Potential Competition, and the FreeRider Problem in Tender Offers, Journal of Law, Economics & Organization 10, 35-62. Franks, Julian, and Colin Mayer, 2001, Ownership and Control of German Corporations, The Review of Financial Studies 14, 943-977. Gaspara, José-Miguel, Massimo Massa, and Pedro Matos, 2005, Shareholder Investment Horizons and the Market for Corporate Control, Journal of Financial Economics 76, 135-165. Grossman, Sanford J., and Oliver D. Hart, 1986, The Costs and Benefits of Ownership: A Theory of Vertical and Lateral Integration, The Journal of Political Economy 94, 691-719. Haely, Paul M., Krishna G. Palepu, and Richard S. Ruback, 1992, Does Corporate Performance Improve After Mergers?, Journal of Financial Economics 31, 135-175. Harrington, Scott E., and David G. Shrider, 2007, All Events Induce Variance: Analyzing Abnormal Returns When Effects Vary across Firms, Journal of Financial & Quantitative Analysis 42, 229-256. Hart, Oliver D., and John Moore, 1990, Property Rights and the Nature of the Firm, The Journal of Political Economy 98, 1119-1158. Hart, Oliver D., and Jean Tirole, 1990, Vertical Integration and Market Foreclosure, Brookings Papers on Economic Activity. Microeconomics 1990, 205-286. Hartzell, Jay C., and Laura T. Starks, 2003, Institutional Investors and Executive Compensation, Journal of Finance 58, 2351-2374. Häussler, Carolin, 2006, When Does Partnering Create Market Value?, European Management Journal 24, 1-15. Hecker, Daniel, 1999, High-Technology Employment: A Broader View, Monthly Labor Review 122, 18-28. Helwege, Jean, Christo Pirinsky, and René M. Stulz, 2007, Why Do Firms Become Widely Held? An Analysis of the Dynamics of Corporate Ownership, Journal of Finance 62, 995-1028.
106
Henderson, Glenn V., 1990, Problems and Solutions in Conducting Event Studies, The Journal of Risk and Insurance 57, 282-306. Hertzel, Michael, and Richard L. Smith, 1993, Market Discounts and Shareholder Gains for Placing Equity Privately, Journal of Finance 48, 459-485. Hertzel, Mitchael, Michael Lemmon, James S. Linck, and Lynn Rees, 2002, Long-Run Performance following Private Placements of Equity, Journal of Finance 57, 2595-2617. Holderness, Clifford G., 2003, A Survey of Blockholders and Corporate Control, Economic Policy Review 9, 51-63. Holderness, Clifford G., and Dennis P. Sheehan, 1985, Raiders or Saviors? The Evidence on Six Controversial Investors, Journal of Financial Economics 14, 555-346. Holderness, Clifford G., and Dennis P. Sheehan, 1988, The Role of Majority Shareholders in Publicly Held Corporations, Journal of Financial Economics 20, 317346. Jenkinson, Tim, and Alexander Ljungqvist, 2001, The Role of Hostile Stakes in German Corporate Governance, Journal of Corporate Finance 7, 397-446. Jensen, Michael C., 1986, Agency Costs of Free Cash Flow, Corporate Finance, and Takeovers, The American Economic Review 76, 323-329. Jensen, Michael C., and William H. Meckling, 1976, Theory of the Firm: Managerial Behavior, Agency Costs and Ownership Structure, Journal of Financial Economics 3, 305-360. Kahle, Kathleen M., and Ralph A. Walkling, 1996, The Impacts of Industry Classifications on Financial Research, Journal of Financial & Quantitative Analysis 31, 309-335. Kale, Jayant R., and Husayn Shahrur, 2007, Corporate Capital Structure and the Characteristics of Suppliers and Customers, Journal of Finance 83, 321-365. Kang, Jun-Koo, and Jin-Mo Kim, 2007, The Geography of Block Acquisitions, Working Paper, Michigan State University. Kirchmaier, Thomas, and Jeremy Grant, 2005, Corporate Ownership Structure and Performance in Europe, European Management Journal 2, 231-245. Klein, Benjamin, Robert G. Crawford, and Armen A. Alchian, 1978, Vertical Integration, Appropriable Rents, and the Competitive Contracting Process, Journal of Law & Economics 21, 297-326. Köke, Jens, 2004, The Market for Corporate Control in a Bank-Based Economy: A Governance Device?, Journal of Corporate Finance 10, 53-80. La Porta, Rafael, Florencio Lopez-de-Silanes, and Andrei Shleifer, 1999, Corporate Ownership around the World, Journal of Finance 54, 471-517.
107
Lawson, Ann M., 1997, Benchmark Input-Output Accounts for the U.S. Economy, Survey of Current Business 77, 36-82. Maksimovic, Vojislav, and Gordon M. Phillips, 2001, The Market for Corporate Assets: Who Engages in Mergers and Asset Sales and are there Efficiency Gains?, Journal of Finance 56, 2019-2065. Malueg, David A., 1992, Collusive Behavior and Partial Ownership of Rivals, International Journal of Industrial Organization 10, 27-34. Maquieira, Carlos P., William L. Megginson, and Lance A. Nail, 1998, Wealth Creation versus Wealth Redistribution in Pure Stock-For Stock Mergers, Journal of Financial Economics 48, 3-33. Martynova, Marina, and Luc Renneboog, 2006, Mergers and Acquisitions in Europe, ECGI - Finance Working Paper No. 114/2006. Mathews, Richmond D., 2006, Strategic Alliances, Equity Stakes, and Entry Deterrence, Journal of Financial Economics 80, 35-79. Maug, Ernst, 1998, Large Shareholders as Monitors: Is There a Trade-Off between Liquidity and Control?, Journal of Finance 53, 65-98. Mikkelson, Wayne H., and M. Megan Partch, 1986, Valuation Effects of Security Offerings and the Issuance Process, Journal of Financial Economics 15, 31-60. Mikkelson, Wayne H., and M. Megan Partch, 1988, Withdrawn Security Offerings, Journal of Financial & Quantitative Analysis 23, 119-133. Mikkelson, Wayne H., and Richard S. Ruback, 1985, An Empirical Analysis of the Interfirm Equity Investment Process, Journal of Financial Economics 14, 523-553. Milanesi, Enzo M., and Alexander Winterstein, 2002, Minority Shareholding, Interlocking Directorships and the EC Competition Rules - Recent Commission Practice, Competition Policy Newsletter 15-19. Morck, Randall, Andrei Shleifer, and Robert W. Vishny, 1988, Management Ownership and Market Valuation: An Empirical Analysis, Journal of Financial Economics 20, 293-315. Morck, Randall, Andrei Shleifer, and Robert W. Vishny, 1990, Do Managerial Objectives Drive Bad Acquisitions?, Journal of Finance 45, 31-48. Mulherin, J. Harold, and Audra L. Boone, 2000, Comparing Acquisitions and Divestitures, Journal of Corporate Finance 6, 117-139. Nenova, Tatiana, 2003, The Value of Corporate Voting Rights and Control: A CrossCountry Analysis, Journal of Financial Economics 68, 325-351. Nocke, Volker, and Lucy White, 2007, Do Vertical Mergers Facilitate Upstream Collusion?, The American Economic Review 97, 1321-1339. Noe, Thomas H., 2002, Investor Activism and Financial Market Structure, The Review of Financial Studies 15, 289-318.
108
Ordover, Janusz A., Garth Saloner, and Steven C. Salop, 1990, Equilibrium Vertical Foreclosure, The American Economic Review 80, 127-142. Park, Yun W., Zekiye Selvili, and Moon H. Song, 2008, Large Outside Blockholders as Monitors: Evidence from Partial Acquisitions, International Review of Economics & Finance 17, 529-545. Pisano, Gary P., 1989, Using Equity Participation to Support Exchange: Evidence from the Biotechnology Industry, Journal of Law, Economics & Organization 5, 109126. Reynolds, Robert J., and Bruce R. Snapp, 1986, The Competitive Effects of Partial Equity Interests and Joint Ventures, International Journal of Industrial Organization 4, 141-153. Russo, Francesco, 2006, Abuse of Protected Position? Minority Shareholdings and Restriction of Markets' Competitiveness in the European Union, Working Paper No. 2006-12, Amsterdam Center for Law & Economics. Salinger, Michael A., 1988, Vertical Mergers and Market Foreclosure, The Quarterly Journal of Economics 103, 345-356. Serra, Ana P., 2004, Event Study Tests: A Brief Survey, Revista Electrónica de Gestao Organizacional 2, 248-255. Servaes, Henri, 1991, Tobin's Q and the Gains from Takeovers, Journal of Finance 46, 409-419. Shahrur, Husayn, 2005, Industry Structure and Horizontal Takeovers: Analysis of Wealth Effects on Rivals, Suppliers, and Corporate Customers, Journal of Finance 76, 61-98. Shleifer, Andrei, and Robert W. Vishny, 1986, Large Shareholders and Corporate Control, Journal of Political Economy 94, 461-488. Shome, Dilip K., and Sudhir Singh, 1995, Firm Value and External Blockholdings, Financial Management 24, 3-14. Smith, Adam, 1776, The Wealth of Nations, New York: Modern Library, 1937. Smith, Richard L., and Joo-Hyun Kim, 1994, The Combined Effects of Free Cash Flow and Financial Slack on Bidder and Target Stock Returns, Journal of Business 67, 281-310. Spencer, Carolyn, Aigbe Akhigbe, and Jeff Madura, 1998, Impact of Partial Control on Policies Enacted by Partial Targets, Journal of Banking & Finance 22, 425-445. Stuart, Toby E., Hoang Ha, and Ralph C. Hybels, 1999, Interorganizational Endorsements and the Performance of Entrepreneurial Ventures, Administrative Science Quarterly 44, 315-349. Sudarsanam, Sudi, 1996, Large Shareholders, Takeovers and Target Valuation, Journal of Business Finance & Accounting 23, 295-314.
109
Warner, Jerold B., Ross L. Watts, and Karen H. Wruck, 1988, Stock Prices and Top Management Changes, Journal of Financial Economics 20, 461-492. White, Halbert, 1980, A Heteroskedasticity-Consistent Covariance Matrix Estimator and a Direct Test for Heteroskedasticity, Econometrica 48, 817-838. Williamson, Oliver, 1979, Transaction-Cost Economics: The Governance of Contractual Relationships, Journal of Law & Economics 22, 233-261. Wruck, Karen Hopper, 1989, Equity Ownership Concentration and Firm Value - Evidence From Private Equity Financings, Journal of Financial Economics 23, 3-28. Martynova and Renneboog (2006)Faccio and Lang (2002) Faccio and Stolin (2006) Lawson (1997)